Professional Documents
Culture Documents
't.-^^fr-
15317
SCHAVM'S OVTLIPiE OF
ADVABf CED
C A L C IJ LIT S
;P!972
BY
MURRAY
R.
SPIEGEL, Ph.D.
Professor of Mathematics
Preface
The
people.
commonly
subject
To some
it
called
To
others
it
advanced topics which are considered important but which cannot be covered
in
an elementary
course.
In this book an effort has been made to adopt a reasonable compromise between these
extreme approaches which, it is believed, will serve a variety of individuals. The early chapters
of the book serve in general to review and extend fundamental concepts already presented
in
elementary calculus. This should be valuable to those who have forgotten some of the calculus
studied previously and who need "a bit of refreshing". It may also serve to provide
a common
background for students who have been given different types of courses in elementary calculus.
Later chapters serve to present special advanced topics which are fundamental to the
scientist,
if
he
is
to
become
This book has been designed for use either as a supplement to all current standard textbooks or as a textbook for a formal course in advanced calculus. It should also prove useful
to students taking courses in physics, engineering or any of the
numerous other fields in which
advanced mathematical methods are employed.
Each chapter begins with a clear statement of pertinent definitions, principles and theorems
together with illustrative and other descriptive material. This is followed by graded
sets of solved
and supplementary problems. The solved problems serve to illustrate and amplify the theory,
bring into sharp focus those fine points without which the student continually feels
himself on
unsafe ground, and provide the repetition of basic principles so vital to effective
learning.
Numerous proofs of theorems and derivations of basic results are included among the solved
The large number of supplementary problems with answers serve as a complete
problems.
Topics covered include the differential and integral calculus of functions of one or more
and their applications. Vector methods, which lend themselves so readily to concise
notation and to geometric and physical interpretations, are introduced early and used
whenever
they can contribute to motivation and understanding. Special topics include
line and surface
variables
and integral theorems, infinite series, improper integrals, gamma and beta functions,
and Fourier series. Added features are the chapters on Fourier integrals, elliptic integrals and
functions of a complex variable which should prove extremely useful in the
study of advanced
engineering, physics and mathematics.
integrals
Considerably more material has been included here than can be covered in most courses.
This has been done to make the book more flexible, to provide a more useful book
of reference
and
wish to take
this
at perfection
M. R. Spiegel
Rensselaer Polytechnic Institute
December, 1962
for
by the author.
CONTENTS
Page
Chapter
NUMBERS
Chapter
20
Functions. Graph of a function. Bounded functions. Monotonic functions. Inverse functions. Principal values. Maxima and minima. Types
of functions.
Special transcendental functions.
Limits of functions.
Right and
left
Continuity.
hand
left
Theorems on continuity.
Chapter
limits.
Right and
SEQUENCES
41
Chapter
DERIVATIVES
57
Definition of a derivative.
tiability
in
an
interval.
Right and
Sectional
left
hand
derivatives.
differentiability.
Differen-
Graphical
inter-
Chapter
INTEGRALS
Definition of a definite integral.
integrals.
Mean value theorems
80
Measure
for
zero.
integrals.
Properties of definite
Indefinite
integrals.
CONTENTS
Page
Chapter
PARTIAL DERIVATIVES
101
Implicit
of
two
Domain
variables.
or
more
functions.
Jacobians.
Theorems on Jacobians.
Mean
Chapter
variables.
of a function.
Partial
Transformations.
value theorems.
VECTORS
134
Laws
Unit
Vectors and scalars.
Components of a vector. Dot or
Rectangular unit vectors.
vectors.
scalar product. Cross or vector product. Triple products. Axiomatic
approach to vector analysis. Vector functions. Limits, continuity and
derivatives of vector functions. Geometric interpretation of a vector
Formulas involving V.
Gradient, divergence and curl.
derivative.
Vector algebra.
of vector
algebra.
Chapter
161
Chapter
MULTIPLE INTEGRALS
Double integrals. Iterated integrals.
180
Triple integrals.
Transformations
of multiple integrals.
Chapter
10
INTEGRAL THEOREMS
195
Evaluation of line
integrals.
Chapter
11
INFINITE SERIES
224
Convergence and divergence of infinite series. Fundamental facts concerning infinite series. Special series. Geometric series. The p series.
Tests for convergence and divergence of series of constants. Comparison
Quotient test. Integral test. Alternating series test. Absolute
test.
and conditional convergence. Ratio test. The nth root test. Raabe's
Gauss' test. Theorems on absolutely convergent series. Infinite
test.
sequences and series of functions. Uniform convergence. Special tests
test. Dirichlet's test.
for uniform convergence of series. Weierstrass
Theorems on uniformly convergent series. Power series. Theorems on
CONTENTS
Page
Double
series.
Infinite products.
Summability.
Asymptotic
series.
Chapter
12
IMPROPER INTEGRALS
260
im-
Chapter
13
285
Chapter
14
function.
FOURIER SERIES
298
Fourier
Chapter
15
series.
Boundary-value problems.
FOURIER INTEGRALS
The Fourier
integral.
Fourier transforms.
convolution theorem.
Chapter
16
Orthogonal functions.
321
ELLIPTIC INTEGRALS
The incomplete
331
the first kind. The incomplete elliptic
incomplete elliptic integral of the third
elliptic integral of
kind.
Chapter
17
345
Limits and continuity. Derivatives. Cauchy-Riemann equations. Integrals. Cauchy's theorem. Cauchy's integral formulas. Taylor's series. Singular points. Poles. Laurent's series. Residues. Residue
theorem. Evaluation of definite integrals.
Functions.
INDEX
373
chapter
Numbers
SETS
Fundamental in mathematics is the concept of a set, class or collection of objects
having specified characteristics. For example we speak of the set of all university
professors, the set of all letters A,B,C,D, .. .,Z of the English alphabet,
etc.
The individual
objects of the set are called members or elements. Any part of a set is called
a subset of
the given set, e.g. A, B,C is a subset of A,B,C,D,..., Z. The set consisting
of no elements
is
called the
empty
REAL NUMBERS
The following types of numbers are already familiar
1.
Natural numbers
to the student.
1,2, 3, 4,
., also called positive integers, are used in counting
of a set.
The symbols varied with the times, e.g. the Romans used
III, IV, ....
The sum a + b and product a-b or ab of any two natural num.
members
I, II,
bers a and b is also a natural number. This is often expressed by saying that
the set of natural numbers is closed under the operations of addition and
multiplication, or satisfies the closure property with respect to these operations.
2.
The
3.
set of positive
is
such as f -f,
arose to permit solutions of
integers a and b where b v^ 0. This leads to the
of multiplication, and we write x-a/b or a-^b
the denominator.
,
The set of integers is a subset of the rational numbers, since integers correspond to rational numbers where & = 1.
4.
numbers such as \/2 and tt are numbers which are not rational, i.e.
cannot be expressed as % (called the quotient of a and b) where a and b are integers
Irrational
and b^Q.
The
set of rational
DECIMAL REPRESENTATION
of
is called
REAL NUMBERS
Any real number can be expressed in decimal form, e.g. 17/10 = 1.7, 9/100 = 0.09,
1/6 = 0.16666 .... In the case of a rational number the decimal expansion either terminates
or, if it does not terminate, one or a group of digits in the
expansion will ultimately
repeat as, for example, in i = 0.142857 142857 142
In the case of an irrational number such as V2 = 1.41423. .. or ,7 = 3.14159... no such repetition can occur.
We can
always consider a decimal expansion as unending, e.g. 1.375 is the same as 1.37500000.
or 1.3749999
To indicate recurring decimals we sometimes place dots over the ree.g. \ = 0.142857, ^ = 3.16.
The decimal system uses the ten digits 0, 1,2,
It is possible to design number
., 9.
systems with fewer or more digits, e.g. the binary system uses only two digits
and 1
(see
Problems 32 and
33).
NUMBERS
GEOMETRIC REPRESENTATION
of
[CHAP.
REAL NUMBERS
of real
-4-3-2-10
-5
in
Fig. 1-1
The
numbers
set of real
is
negative.
many
(or irrational)
OPERATIONS
with
REAL NUMBERS
R of real
belong to R
2.
3.
4.
5.
6.
7.
(or irrational
is
call
set.
numbers, then:
+ b and ab
a+b = b+a
a + {b + c) = {a + b) + c
ab = ba
a(bc) (ab)c
a{b + c) ab + ac
= + a = a, l'a = a*l =
a +
a
numbers) on the
This leads us to
Closure law
a
to addition, 1
with
respect
the
identity
called
is called
respect to multiplication.
8.
X
9.
set,
is
is
denoted by a.
is
denoted by
In general any
INEQUALITIES
If a 6 is a positive number we
2.
3.
4.
5.
and
or
1 >
2;
a;
means that
a;
is
Law of trichotomy
Law of transitivity
CHAP.
NUMBERS
1]
ABSOLUTE VALUE
of
REAL NUMBERS
a<0, and
if
Examples:
1.
\ah\
2.
|a
+ 5|
\a-b\
3.
real
number
|-||
a,
|+2|
5,
\a\
+
-
|,
|-V2|
\ahc...m\
or
l&l
=
|a + 5 + c +
or
l&l
|a|
2,
\h
^2,
defined as a
is
|a|,
|o|
+m| S
if
a>
0,
-a
if
o.
|w|
|a|
+ |6| +
|c|
+ |m|
\h\
denoted by
= 0.
|-5i
|a|
(real
a\.
is
EXPONENTS
and ROOTS
The product a-a. .a of a
called the exponent and a is
.
is
1.
to
a"
any
real
-a''
real
cluded.
number a by
In particular by using
definitions a" = 1, a~ = l/a".
p times
itself
The following
0"+'
3.
(a")''
is
rules hold.
a"''
numbers are possible so long as division by zero is ex2, with p=^q and p =
respectively, we are led to the
may
If
p and q are
we
positive integers,
define
a''^"
= \^.
LOGARITHMS
If
a''
= N,p
is
lOgaMA^
= logaM + logaA^
3.
logaAf
2.
loga
= logaM-logaiV
= rlogaM
In practice two bases are used, the Briggsian system uses base a
tem uses the natural base a = e = 2.71828 ....
AXIOMATIC FOUNDATIONS
The number system can be
of the
= 10,
up
built
plication (although it is possible to start even further back with the concept
of sets),
find that statements 1-6, Page 2, with
as the set of natural numbers,
we
do not hold.
and
Taking 7 and 8 as additional requirements, we introduce the numbers -1, -2, -3,
Then by taking 9 we introduce the rational numbers.
0.
NUMBERS
[CHAP.
Operations with these newly obtained numbers can be defined by adopting axioms 1-6,
R is now the set of integers. These lead to proofs of statements such as (-2) (-3) = 6,
-(-4) = 4, (0)(5) = 0, etc., which are usually taken for granted in elementary mathematics.
where
can also introduce the concept of order or inequality for integers, and from these
For example if a, b, c, d are positive integers we define
eld if and only if ad > be, with similar extensions to negative integers.
We
a/b
>
Once we have the set of rational numbers and the rules of inequality concerning
them, we can order them geometrically as points on the real axis, as already indicated.
We can then show that there are points on the line which do not represent rational numbers (such as v^,7r,etc.). These irrational numbers can be defined in various ways one
From this we can show that
of which uses the idea of Dedekind cuts (see Problem 34).
the usual rules of algebra apply to irrational numbers and that no further real numbers
are possible.
is
called a one-dimensional
point set.
The set of points x such that a^x^b is called a closed interval and is denoted by
The set a<x<b is called an open interval, denoted by {a,b). The sets a<x^b
[a,b].
and a g X < b, denoted by {a, b] and [a, b) respectively, are called half open or half closed
intervals.
The
|x|
<
4,
i.e.
-4 <
a:
< 4,
of a set,
is
is
called a variable.
interval.
The set x>a can also be represented by a<x<cc. Such a set is called an
or unbounded interval. Similarly - < x < represents all real numbers x.
infinite
COUNTABILITY
denumerable
respondence with the natural numbers.
set is called countable or
Example:
2, 4, 6, 8,
if
its
is
ence shown.
Given set
Natural numbers
set is infinite if it
infinite set
The
or
which
is
8...
i;
...
itself.
An
countable
set of rational
all real
numbers
is
set which is
of elements in a set is called its cardinal number.
countably infinite is assigned the cardinal number K^, (the Hebrew letter aleph-null). The
with this set)
set of real numbers (or any sets which can be placed into 1-1 correspondence
the continuum.
is given the cardinal number C, called the cardinality of
The number
CHAP.
NUMBERS
1]
NEIGHBORHOODS
The
of the point
is called
The
a.
a deleted
neighborhood of
|a;
neighborhood
is
called a
in
which
=a
a;
is
excluded,
a.
LIMIT POINTS
8 > 0, however small, we can always find a member x of the set which is not equal
but which is such that \x-l\ < 8. By considering smaller and smaller values of 8
see that there must be infinitely many such values of x.
for any
to
we
A
point.
finite set
has infinitely
many
An
may
infinite set
or
may
numbers
limit points.
The
is
is
set of rational
not a
member
numbers
of the set
BOUNDS
If for all
above and
numbers
is
a;
of a set there
M such
a number
Similarly if
is
an upper bound.
called
that
x^m,
x^M,
the set
is
m^x^M,
x we have
the set is called bounded.
a number such that no member of the set is greater than
but there is at
least one member which exceeds
for every e > 0, then
is called the least upper
bound (l.u.b.) of the set. Similarly if no member of the set is smaller than
but at least
one member is smaller than
+ e for every >0, then
is called the greatest lower
bound (g.l.b.) of the set.
is
If
If for all
is
M-
WEIERSTRASS-BOLZANO THEOREM
The Weierstrass-Bolzano theorem states that every bounded infinite set has at least
one limit point. A proof of this is given in Problem 23, Chapter 3.
aox"
where
ttix"-'
a^x"-^
...
equation
an-ix
{1)
the equation,
ai, aa,
is
Examples
called
2 =
and
x^
2 =
The numbers
and e can be shown to be transcendental numbers.
determine whether some numbers such as e-n- or e + tt are algebraic or not.
-n-
The
set of algebraic
of transcendental
set (see
Problem
We
still
cannot
NUMBERS
The
[CHAP.
x^
+1 =
We
can consider a complex number as having the form a + bi where a and b are
called the real and imaginary parts, and i \/ 1 is called the imaginary
unit. Two complex numbers a + bi and c + di are equal if and only if a = c and b = d. We
can consider real numbers as a subset of the set of complex numbers with b 0. The
complex number + Oi corresponds to the real number 0.
real
numbers
bi is defined as a
of a
bi is defined as |a
bi\
bi.
y/aF+~b^.
of the
The com,plex
complex number
The set of complex numbers obeys rules 1-9 of Page 2, and thus constitutes a field.
In performing operations with complex numbers we can operate as in the algebra of real
numbers, replacing t^ by 1 when it occurs. Inequalities for complex numbers are not
defined.
From
POLAR FORM
of
COMPLEX NUMBERS
Y
*P(3,4)
Q(-3,3)
Pio'.y)
T(2.5, 0)
-t-
R(-2.5, -1.5)
*S(2,
-2)
Y'
Y'
Fig. 1-3
Fig. 1-2
Since a complex number x + iy can be considered as an ordered pair (x, y), we can
represent such numbers by points in an xy plane called the complex plane or Argand
diagram. Referring to Fig. 1-3 above we see that x p cos
y = p sin ^ where
p = \/x^ + y^ = \x+iy\ and ^, called the amplitude or argument, is the angle which line
OP makes with the positive x axis OX. It follows that
<l>,
ly
p(cos
(j>
cis
+
p
sin ^)
and
^ instead of cos
<^
<j>
(2)
CHAP.
NUMBERS
1]
If
Zi
we can show
Xi
iyi
pj(cos
+ i sin
<j>^
and
<j>^)
Z2
X2
+ iy^ =
Pii'ios
<i>^
sin ^2)
that
Z1Z2
^{cos(</.i-,^2)
3"
{p(cos<^
+ ^2) +
*'
sin
(</>!
+ .^2)}
(S)
*sin(,^j-<^2)}
isin^)}"
p" {cos
kV)
714,
sin
n<j,)
(5)
where n
(p(cos^
p-|cos(i^)+^sin(i^)}
isin^)}*''"
{6)
A.
= 0,l,2,3,...,n-1
from which it follows that there are in general n different values for z^"^.
will show that e** = cos
+ i sin ^ where e = 2,71828
This
we
<j!>
called
Euler's
formula.
MATHEMATICAL INDUCTION
The
integers.
It is especially useful in
following steps.
1.
2.
Assume
3.
integer).
positive integer.
From
This
4.
l (or
is
Since the statement is true for n = l [from step 1] it must [from step 3] be true
for w = 1 + 1 = 2 and from this for = 2 + 1 = 3, etc., and so must be true for
all
positive integers.
Solved Problems
OPERATIONS
1.
NUMBERS
with
p(gr),
(d) (pq)r,
+ (2/ + z) =
+ y) + z =
(a)
a;
(6)
(x
{c)
p{qr)
(d)
(pq)r
The
(4
=
=
x(v
= -J, and
(6)
and
(d)
= |,
evaluate
(a)
{y
+ z),
(&) {x
+ y) +
z,
+ 12 =
16
19 - 3 = 16
are equal illustrates the associative law of addition,
|{(-i)(|)} = (|)(-^) =
{(f)(-i)}(|) = (-fs)(f) =
(c)
+ q).
+ [IB + (-3)] =
+ 15) + (-3) =
(e)
(f)(-i)
(-l)(|)
= -^ = -3L
= -I = -3L
a;(p
distributive law.
(4)(-i)
= |-| =
|-|=f=2
using the
NUMBERS
2.
why we do
Explain
If
we
Ox
= 0.
(a)
However, this
we
represent,
(6)
As
not consider
in (a), if
is
consider
we
number
it
(a)
(b)
[CHAP.
as numbers.
(if it exists)
true for
all
undefined.
number x
(if it exists)
such that
Oa;
= 1, we
no such number.
is
3.
Simplify
~2
a;
# 3.
For
RATIONAL
4.
ZTq
=3
a;
and
Qw XT\
XT
4rw*
3)
is
not zero,
i.e.
IRRATIONAL NUMBERS
+ 4m =
2(2?n^
+ Im),
is 1
is
= 2m.
Substituting p = 2m
be odd.
Thus p
in p"
= 2q^
yields q^
common
By
whose square
6.
(a;
undefined.
is
5.
^,_^^_3
= 2m*,
so that q^ is even
and q
even.
is
is 2.
Show how to find rational numbers whose squares can be made arbitrarily close to 2.
We restrict ourselves to positive rational numbers. Since (1)* = 1 and (2)* = 4, we are led to choose
rational
Since
= 2.25, we
1.5, e.g.
1.41,
1.42, ...,1.49.
Continuing
(1.414213562)*
7.
in
is less
is
approximations,
this
a root
we
OOP"
where
is to
ao, ai,
ao exactly.
aip"-'q
a2p''-*q*
or dividing by p,
OOP"-'
e.g.
dip-'-^g
a,-ipg-'
ttn-ig"-'
ag
q",
the result
(1)
-^^
Since the left side of (2) is an integer the right side must also be an integer.
p does not divide g" exactly and so must divide a.
(z)
Then
since
p and g are
relatively prime,
must
8.
divide
first
term of
{!)
and dividing by
g,
we can show
that g
ao.
CHAP.
9.
NUMBERS
1]
a and
assume a <
this
5^^ <
<
<
<
6.
a = and
is
= -
where
Then
to both sides, 2a
26 and
6.
.
6.
^r
To prove that
97^0,8^0.
is
Then by adding a
6.
Thus
numbers, then
6 are rational
To prove
is
^^
p, g, r, s are integers
and
rational number.
INEQUALITIES
10.
11.
is
ic
a;
a;
cc
3(2
x) S
a;
6-43
+ 6-3a;g4-a;, 6-2a;g4-a;,
when
a;
is
x^
- Bx -
<
10
- 2x
2aa;,
-a;,
2a;
i.e.
a;
g 2.
a;''-3x-2-10 +
<
2a;
0,
a;'
- a;
12
<
or
(a;
- 4)(a; + 3) <
Case 1
a; - 4 >
and a; + 3 < 0, i.e.
both greater than 4 and less than 3.
:
Case
12. If
>
a;
and x
<
-3.
This
is
and a; + 3 > 0, i.e. a; < 4 and x > -3. This is possible when -3 <
2: a; - 4 <
Thus the inequality holds for the set of all x such that 3 < a; < 4.
ago
A
and
&
0,
method of proof
+ b) g
a;
<
4.
t/oE.
In this problem we start with the required result to obtain successively a + 6 S 2^/ab, {a + by g
or a'-2ab + b'' S 0, i.e. (a-bf g 0, which is known to be true. Retracing the steps, the
4ab
result follows.
Another method:
Since
we have a 2^/ab +
(Va-y/bY S
o or
g ^aiUi
A(a+6) ^ -/^,
negative. The left and right sides are called respectively the arithmetic
of the numbers ai,
On.
.
13. If
Ui, a2,
For
all real
6i, bi,
a2b2
numbers
X,
are non-
.,b
anb^f
(a?
a|
al)(bl
6|
+ &')
we have
(aiX+6i)'
Expanding and
ai,
and
(aibi
where
(aiX
biV
(a\
6)^
A'\'
+ 2C\ +
B"
(1)
where
A^
Now
{1)
a?
+ oi +
+ a,
B'
= bl+bl+
+bl,
C =
oi6i
+ 0262 +
+ o6
{S)
can be written
X^
gx
or
(x
+^y
But this last inequality is true for all real X if and only it
gives the required inequality upon using (2).
+ 1; -
-^ - -^ ^
A
A
(S)
or
C^S A^B'
which
NUMBERS
10
14.
111
+ +
Prove that
Let
15.
'
'^
"
'
2r^ -
^^
^+...+^+^
Subtracting,
1S
= i -
-^-
>
1.
+^
^+
IS.
positive integers
^^^
= i + 1 +
Then
EXPONENTS, ROOTS
...
Thus S
g^
==
for
all n.
LOGARITHMS
and
=-
(a)
aF^^P^
(c)
log2/3(f)
(log 6)(logb a)
Since
>
ilf
AT
0,
Let log
i^
>
MM
77
X, log
""
^.^^.^ ^
^ ^Z,^^,^ ^
or
(f)"'
a;
-3.
2/
5. 10-= or 0.00005
and
a, 6 7^ 1.
b, b"
a>0
and
= f - ()' =
loga b = x, logb a =
= a and m = xy.
(a^) = a^ = 6 = a we have
Then a'=
16. If
(f^
Let
u.
q-2
^W.lO^
Then
X.
Q4 + 8-14
(^)
(d)
N = y.
-
""
_"
= a'
a^l,
but
Then a'-M,
a"
a""
or
=1
prove that
a'^N
loga
loga
M - loga
A^.
and so
a;^/
lo&]v
= logaM -
loga AT
COUNTABILITY
17.
"'"
Sn
[CHAP.
numbers between
and
1 inclusive is countable.
plished as follows.
Rational numbers
Oliifilif
Natural numbers
123456789...
18. If
and
or
is
countable, there
Suppose elements of
or
is
or
elements of
are
countable since
is
Case
1 inclusive is countable
all
elements from
Since A
bers so that
from
and
...
by
all distinct
we can
Natural numbers
ai
61
^
^
A
V
as
62
az
63
**
^
^
"^
...
...
Case 2: If some elements of A and B are the same, we count them only once as in Problem 17.
the set of elements belonging to A or B (or both) is countable.
Then
CHAP.
NUMBERS
1]
The
and B, denoted by
A UB
set consisting
or
union of
A+B.
or
19.
which belong to
11
all
or
numbers
is
countable.
is
1.
i.e.
From Problem 18 it then follows that the set consisting of all positive rational numbers is
countable, since this is composed of the two countable sets of rationals between
and 1 and those
greater than or equal to 1.
From
20.
this
real
number
0, 1,2,
the set of
in
rational
all
numbers
in
numbers
non-countable.
is
[0, 1]
[0, 1]
is
mi aiUs
where
oi, as,
are any of
.,9.
We
assume that numbers whose decimal expansions terminate such as 0.7324 are written
and that this is the same as 0.73239999
If all real numbers in [0, 1] are countable we can place them in 1-1 correspondence with the
natural numbers as in the following list.
0.73240000...
We now
<^
O.aii
an
ais
an
->
O.Uii
a-zz
023 a2i
(r->
O.asi
ttsa
ass 034 ..
form a number
0.6i 62 63 64
where
61
# an,
62
# 022,
63 f^ ass, 64 v^ 044,
and where
all
b's
This number, which is in [0, 1], is different from all numbers in the above
the list, contradicting the assumption that all numbers in [0, 1] were included.
list
and
is
all 9's.
thus not in
Because of this contradiction it follows that the real numbers in [0, 1] cannot be placed in 1-1
correspondence with the natural numbers, i.e. the set of real numbers in [0, 1] is non-countable.
infinite set of
least
is
(d) Is
set.
Since
2
is
We
(6)
all
members
numbers
and greatest
(l.u.b.)
is
(namely
member
1)
of the set
which exceeds
member
is
bounded,
(g.l.b.)
(b)
of the set.
(e)
How
Determine the
(c) Prove that
illus-
bounded;
a lower bound.
Since no
bound
theorem?
an upper bound, 1
1, i, i, i,
lovs^er
THEOREM
is
(e.g. |
(e.g.
^).
greater than 1 and since there is at least one member of the set
for every positive number e, we see that 1 is the l.u.b. of the set.
Since no
NUMBERS
12
(c)
we
we
small
take
we
see that
is
a positive
set.
To put this another way,
always includes members of the set, no matter how
> 0.
The
(e)
Since the set is bounded and infinite it must, by the Weierstrass-Bolzano theorem, have at least
one limit point. We have found this to be the case, so that the theorem is illustrated.
set.
TRANSCENDENTAL NUMBERS
and
23.
<\x\<
set.
(e.g.
neighborhood of
(d)
ALGEBRAIC
22.
any deleted
see that
[CHAP.
is
it
numbers
algebraic
all
a countable
is
which
is
set.
Let
+ '\/3,
an algebraic number.
\ao\
oo a;"
!""'
...
|ai|
|a|
n.
number
number
COMPLEX NUMBERS
24.
(d)
(- 2
+ 5)i = -2 + 3i
- 4i) = - 7 + 3i - 2 + 4i = - 9 + 7i
(3 - 2i)(l + 3i) = 3(1 + Si) - 2i{l + 3i) = 3 + 9i - 2i - 6P = 3 + 9i - 2i +
-5 + 5i _
-5 + 5i 4 + 3i _ (-5 + 5z)(4 + 3i) _ -20 - 15i + 20i + 15t^
16 + 9
4-3i
4 - 3i
4 + 3i
16 - 9i'
-35 + 5t
5(-7 + i)
-7,1.
-5- + 5'
25
25
(- 7
(e)
+ 3i) -
5i
(2
+ i' + i^ + i^ + i^
1 + i
1
1
(/)
- 4il
|4
1
iff)
25. If zi
+ 3i\
3i
3t
and
Let
\ziZ2\
Zi
Z2
+ 3i
-6i
2^
25
= V(or+(-j%r = f
\ziZ2\
+ iyz.
+ iyi){x2 + iy2) = xiXi yiy2 + i{xiy2 + Xzyi)
= ^/{XiXz yiyiY + (xiy^ + XtVif = ^/x\xl + y\yl + ^Ivl +
Xi
(5)(5)
10
di"
1,1.
i+1
VW+iFWVW+W
iyi,
Z2
\/ (.x\
+ y\){x\ +
y\)
\zi\ \Z2\.
Then
X2
(xi
_ i-1-i+l + i
(V)H
1-
3i
+ (V-f +
+i
i i^
(i')(j)
\fx\
y^ \/x\
y\
\xi
iyi\\xi
^Ivl
iy2\
|l| \Z2\.
7i
CHAP.
26.
NUMBERS
1]
Solve
- 2x -
a;3
13
0.
The possible rational roots using Problem 7 are 1, 2, 4. By trial we find x = 2 is a root.
Then the given equation can be written (x 2)(x^ + 2x + 2) = 0. The solutions to the quadratic equation
-2
ax"
bx
'^
^4-8
= ~^ ~
"
cb
The
of
2i
1 +
- 4(ic
V6^
2a
= -1
set of solutions is 2,
POLAR FORM
27.
o^"
are
(^
-2 ^/-i _ -2
a=l,
For
= 2,
=2
this
gives
i.
i,
COMPLEX NUMBERS
(a)
3i,
- 1 + ^i,
(6)
- 1,
(c)
-2 -
(d)
2v/3 i.
>y
180
45
-2^3
(a)
(c)
Fig. 1-4
(a)
Amplitude
3
(&)
3i
Amplitude
=
=
45
120
ff/4
p(cos
-1 +
(c)
Amplitude
sin 0)
tt
Amplitude
240
-2 -
28.
Evaluate
(a)
(a)
(-l
(6)
(-1
By Problem
-1 +
47r/3
2\/3
=
=
=
l(cos
[2(cos2r/3
1024 [cos
(2,7/3
1024(-l
= V2 (cos 135 +
J-/3
6,7)
\/(-l)^
Then
= 1. Then
= e"
V(-l)^+(0)'
cis
77
V(-2)*
sin 4,7/3)
+ (V^)'' = yT = 2.
= 26^'^""
2 cis 2,7/3
+ (-2Vsy = 4.
cis 4,7/3 = 4e*''*"
Then
i^'K
= 0, 1, 2
2"'(cos 20,7/3
(2W3 +
[eos
6,r)]
+
=
sin 20,7/3)
1024(cos 2,7/3
sin 2,7/3)
fc '
360)
sin (135
k 360)]
(i5^1+_fc^3601^
135
i
sin
fe-360
0]
are
sin
= -512 + 512^3 i
1)
sin 135)
+ ^)V3 = (V2)-
fe
sin w)
sin 2,7/3)]"
results for
(-1
+
The
sin 27r/3)
p
3\/2e"'*
Then
(-l
J7-
4(cos 4j7/3
(&)
Modulus
radians.
+ \/3 zy,
27(6)
VSi)^"
Modulus
radians.
jr/4
Modulus p
2(cos 27r/3
1 =
(d)
3\/2(cos
2ir/3 radians.
s/Zi
180
Modulus
radians.
repetitions of these.
in the
complex
Fig. 1-5
a;
NUMBERS
14
[CHAP.
MATHEMATICAL INDUCTION
29.
Prove that
1^
The statement
2^
is
3^
4^
%=
true for
Adding
(k
+ 1)^
P + 2^+ 3^+
1, it
+ n^ =
1 since
n=
for
2^
+ l){2n + 1).
^(1)(1
1)(2
1)
1.
Then
k.
3^
1"
in{n
A;^
:^k{k
l){2k
1)
to both sides,
+ l){2k+l) + (k + iy = {k + l)[^k{2k + 1) + k + 1]
+ l)(2fc^ + 7fc + 6) = ^(fe + l)(fc + 2)(2fe + 3)
statement is true for n = k + 1 if it is true for n k. But since it is true for
it is true for n 1 + 1 = 2 and for m = 2 + 1 = 3,
., i.e. it is true for all
+fc^
n=
follows that
(fc
+ l)^ =
=
^k{k
|(fe
positive integers n.
30.
Prove that
a;"
2/"
has x
as a factor for
x^ y^ =
n=l
since
The statement is true for
Assume the statement true for n = k,
x y.
assume that x^
i.e.
positive integers n.
all
a;''(a;
i/) +
i;(a;'^
y^
has
Consider
as a factor.
y^)
y&s,
The first term on the right has a; j/ as a factor, and the second term on the right also has x
a factor because of the above assumption.
Thus
Then
x
31.
0!''+'
has
2/''+'
since x^
y^
a;
as a
y&s
2/
has x
(1
has x
(1
But
true for
+ x)""-^ >
is
true for
as a factor, x^
y^
has
is
+ nx
for
positive since x
>
1).
if
x>
-1, Xy^O.
Then we have
= 1.
is true for n
2, it
integers greater than or equal to 2.
(1
n = 2,3,
1 + 2a; + a;^ > 1 + 2a;.
(l + x)^
i.e., {1 + xY > 1 + kx.
+ a;)" >
for
y^
as a factor, etc.
+
2
+ l)x
(fc
+1 =
(1
S,
+ a;)" S
...
and
thus
is
nx
true
is
1,2, 3,
MISCELLANEOUS PROBLEMS
32.
Prove that every positive integer P can be expressed uniquely in the form
+ ai2"-i + 022""^ + + a where the a's are O's or I's.
Dividing P by 2, we have F/2 = (10 2"-' + ai2"-^ + + -i + a/2.
Then
a is the remainder,
or
1,
P =
ao2"
obtained
By continuing
33.
we
unique.
in this
manner,
in the
all
form of Problem
coefficients
32.
can be arranged as follows.
2^23
2 ^11
2)5
12
2 )3
2
Remainder
Remainder
Remainder
Remainder
Remainder
1
1
1
1
and
CHAP.
NUMBERS
1]
The
coefficients are
34.
10 111.
is
Check: 23
2"
2'
2^
1.
Dedekind defined a
cut, section
numbers
15
or partition in the rational number system as a separatwo classes or sets called L (the left hand class) and
(the right hand class) having the following properties:
into
I.
II.
III.
to each class).
It is possible
What
(a)
a smallest number in R.
It is possible
What
(d)
L and
It is possible
What
(c)
in
Let a be the largest rational number in L, and b the smallest rational number in R.
a b or a<b.
We
number
cannot have a
R.
=b
since
by
number
in
is less
Then
either
than every
in
We cannot have a<b since by Problem 9, ^{a+b) is a rational number which would be
greater than a (and so would have to be in R) but less than b (and so would have to be in L), and
by definition a rational number cannot belong to both L and R.
(b)
As an
(c)
As an
(d)
As an
all
rational
R contains
A similar
less
'
NUMBERS
16
[CHAP.
Supplementary Problems
OPERATIONS with NUMBERS
35. Given x = 3, y 2, z = 5,
(a)
yK3y +
{2x
Ams.
(a)
2200,
Find the
36.
^"^^
(a) 2,
RATIONAL
3(2a;
(6) 6,
g^fe
_i
(6)
1)}
= 1,
and
-51/41,
(c)
a;-2
evaluate:
'
2a^6^
+1
^^^^
'
-4,
(aj/
fexf
2(2a;
1)
+ 2) -
12(a;
(c)
Va;''-2a;
-1,
-1
(d)
1,
+ (x-y)(y-z)
+
.^_,X_
{z-x)(x-y)
62/)^
f, _
.
,,,_
(2/-z)(-a!)
'
'
all steps in
each case.
\/SM^8^+T - V2^T2 =
(<^)
(c)
(a;
(d) 1
,_J,_,
Prove that
37.
(6) 32,
- 2) +
8-a;
Ans.
2z),
4{(x
(a)
z)(5x
=|
+5
a;
giving restrictions
any.
if
IRRATIONAL NUMBERS
and
38.
39.
Show that a fraction with denominator 17 and with numerator 1,2, 3, ...,16 has 16 digits in the
repeating portion of its decimal expansion. Is there any relation between the orders of the digits in
these expansions?
40.
Prove that
(a)
V^,
41.
Prove that
(a)
y/E
42.
43.
44.
2*'
Ans.
45.
46.
If
(a)
is
^/^ + y/h
Sa;"
(6)
0,
8/3,
(6)
Ams.
(6) \/5.
(h) yf2
yfi,
18
-2, 3/2
3,
0.428571, (6)2.2360679...
(a)
(6) ^f2
- 9a; +
5a;''
(a)
-2
+ 4a;' - 35a; + 8 =
(c)
"v/s
|(5
from 7 by
differs
+ h^fm =
+ V3 + Vs ^ 12^5 - 2^15 + 14^3 - 7
^^
l-\/3 + V5
(c)
0,
a;*
l(-5
^/ri),
-21a;'
than .000001.
+4 =
0.
y/vf)
+ d-\fm
less
if
and only
if
and
= d.
Prove that
INEQUALITIES
47.
Find the
^"^
X
Ans.
+
(a)
48.
Prove
49.
50.
If
51.
If
52.
53.
Show
.
54.
a'
(a)
|a;
+ 6' =
a;
j/|
all
and
all
|a;|
\y\,
real x, y,
.,n
set of values of
c''
d'
a;"+'
a'0,
z,
24,
x^
1,
(c)
4,
(6)
|x
la;
+ 2| <
<
3/2,
+ y + ] S
+ y^ + z' S
prove that
+ ^^ >
\a
(c)
+ l/a\ a
x"
-{
xy
(d)
ja;]
(d)
la;-5|,
a;
>
\y\
>
^^
3,
^
a;
\z\,
(c)
|a;-2/|
or
a;
< -2
|a;|
|2/|.
+ yz + zx.
ac+hd S
where n
1.
is
any
positive integer.
2.
+ 02 + as) s
y/aiaiaz.
if
and only
if
Op
= kbp,
1,2, 3,
CHAP.
NUMBERS
1]
EXPONENTS, ROOTS
55.
Evaluate
Ans.
(o) 64,
56.
Prove
57.
Prove
(a)
6'''b<'
(6)
(6) 7/4,
loga
LOGARITHMS
and
4'-.
(a)
I7
log.. s (if,),
50,000,
(c)
MAT =
(d) 1/25,
M + Iog
Iog
a giving restrictions
iV,
if
-^/iM^|||M
(c)
()
(_l)./3
(_27)-/'.
-7/144
(e)
(6)
(d)3--3^
log
r loga
any.
COUNTABILITY
58.
(a)
-5 ^
Ans.
59.
(a)
-3.
(b)
(6)
is
and
What
all
rational
numbers
countable.
is
(6)
What
is
the cardinal
number
of the set
in (a)?
GO.
61.
The
intersection of
belonging to both
numbers,
(6) all
irrational
numbers
is
is
non-countable.
all
elements
so is their intersection.
62.
63.
Prove that the cardinal number of the set of points inside a square is equal to the cardinal number
(a) one side, (6) all four sides, (c) What is the cardinal number in this case?
(d) Does a corresponding result hold for a cube?
Ans. (c) C
is
countable.
LIMIT POINTS.
64.
Given the set of numbers 1, 1.1, .9, 1.01, .99, 1.001, .999, .... (a) Is the set bounded?
(6) Does the set
have a l.u.b. and g.l.b.? If so, determine them, (c) Does the set have any limit points? If so, determine
them, (d) Is the set a closed set?
Ans. (a) Yes
(6) l.u.b. = 1.1, g.l.b. = .9
(c) 1
(d) Yes
65.
Given the set -.9, .9, -.99, .99, -.999, .999 answer the questions of Problem
Ans. (a) Yes
(6) l.u.b. = 1, g.l.b. == -1
(c) 1, -1
(d) No
66.
67.
(a)
(6)
68.
Let
no limit points.
be the set of
all
numbers
rational
(6) Is
having denominator
in (0,1)
and
n =
1,2,3,
(a)
Does S
is
not bounded.
(6)
TRANSCENDENTAL NUMBERS
%/3
Prove that
2",
closed?
(a) Give an example of a set which has limit points but which
the Weierstrass-Bolzano theorem? Explain.
ALGEBRAIC
70.
(6)
64.
(a)-
-1/2
(6) -^2
V3 + V2
+ y^ + -/B
71.
72.
is
number
is
+ IT3?'
<^Hm^
Ans.(a)l-4i,
{b)
-9 -
2-4i
(^)
46i,
(1
(f.
+ 7i
{c)^-U,
^^'
- Si) - 3(-2 + i) +
+ i){2 + 3i)(4 - 2i)
{l + 2iril-i)
(a) 2(5
(d)
-1,
(e)
if,
5{i-
(f)f-U
3),
(6)
(3
2^)^
(c)
-41
NUMBERS
18
74.
and
If zi
za
= M,
11
(a)
\Z2\
75.
Prove
76.
Find
77.
|zi|
solutions of
all
(c)
l^sl,
3,
|zi
- 2| ^
|zi|
|z2|.
-li
i,
and
zi
zi
is
two dimensions.
in
78.
79.
Express
Ans.
form
in polar
6 cis
(a)
Evaluate
Ans.
81.
\z^\'
be represented by points Pi and Pa in the Argand diagram. Construct lines OPi and OP2,
the origin. Show that zi + 22 can be represented by the point Pa, where OPs is the diagonal
of a parallelogram having sides OPi and 0P%. This is called the parallelogram law of addition of
complex numbers. Because of this and other properties, complex numbers can be considered as vectors
Let
where
80.
\z\\
(6)
\Zl\
+ Z2I ^
(a) |zi
[CHAP.
-5V2 +
(a)
Determine
Ans.
(c)
(c)
2 cis 57r/3
sin25)][5(cosll0
l-^fli,
(d) 5,
id) 5 cis
sin 110)],
-5i.
(e)
5 els 37r/2
(e)
12 cis 16
^.^^^o^^g cis 62)
(6)
^g
'
-2i
(6)
5-v/2i,
all
(4V2 + 4V2i)'',
(a)
[2(cos25
(a)
-2-2i,
(6)
(6)
7r/6
+ 3t,
3\/3
{a)
(e)(-Jz-iY'\
(6) (-l)''^
=-1,
180
(c)
83.
If
+ Vsi
Prove that 1
324
(6)
82.
i"\
(d)
an algebraic number.
is
zi = picis^i and Z2
Interpret geometrically.
prove
PjCis^j,
(a) Z122
+ ^j).
^1/^2
(&)
s^a)-
MATHEMATICAL INDUCTION
Prove each of the following.
(2m
- 1)
84.
85.
7^^
+
+ -p^ +
l-33-55'7
86.
(a
2'
2(5)^
90.
1(5)
91.
x'"-^
92.
(cos0
93.
A +
sin
a;
2/^"-'
cos
a;
^n[2a
n{n + l){n +
^lljZ-i
*
by
cosw^
cos 2x
2i+l
+ (n- l)d]
4(n
2)
+ l)(n + 2)
^^1
= In^n + ly
3(5)
+ l)
(2w-l)(2w
ar"--^
is divisible
sin 2a;
3'4'5
sin0)"
+ 2d) +
ar"
3'
-^
94.
{a
2-3'4
ar
1^
+ d) +
89.
l'2-3
88.
sin
a;
(5)-^
+
i
cos
nx
j/
for w
sinn^.
nx
=
=
^g
1, 2, 3,
Can
this be proved if
r^
+^x
cos Ax cos
2 sm
is
a rational number?
i)x
sin {n
Z sin
.
(ti
a;
7^ 0, 2ir, 4n-,
+ A)a;
,
-.
ia;
a;
where
is
C,
defined as
. ,
nC
1.
This
is
The
# 0, 2jr, 4jr,
(a+6)
95.
.Cn-iaft""'
Here
p!
Co = 1,
coefficients
Cr is also written
6"
=p(p-l)-l
Ci = n,
nCa
and
,
nin
g-j-
0!
,.
1)
CHAP.
NUMBERS
1]
19
MISCELLANEOUS PROBLEMS
96.
Express each of the following integers (scale of 10) in the scale of notation indicated: (a) 87
64 (three), (c) 1736 (nine). Check each answer.
Ans. (a) 1010111, (6) 2101, (c) 2338
(two),
(6)
97.
If a
Ans.
98.
number
(a)
5,
what
is
number
the
and
can be expressed
(a) 2,
(6)
in the
8?
form
^2"
2"
in the scale of
where the a's can be determined uniquely as O's or I's and where the process may or may not terminate.
The representation O.ai a^
a
is then called the binary form of the rational number.
[Hint:
Multiply both sides successively by 2 and consider remainders.]
.
99.
100.
101.
In
number
what
106.
(c)
(6) 3,
0.2 or 0.2000.
(d)
8,
.,
(c)
What
10.
0.5252.
is
the
number
Show
.,
(d)
number
0.6666...
Ans. 3.28125
two additional symbols t and e must be used to designate the "digits" ten and
Using these symbols, represent the integer 5110 (scale of ten) in the scale of 12.
the
Ans. 5
12?
first digit,
105.
(a) 2,
scale of notation is
eleven respectively.
Ans. 2 e 5 <
104.
103.
new number
is
one-third as large.
is
1.636363
Ans. 18/11
field.
(a) (-3)(0)
number whose square is less than 2, show that x + {2- x")/10 is a larger such
a rational number whose square is greater than 2, find in terms of x a smaller
rational number whose square is greater than 2.
number.
108. Illustrate
(6)
If
is
(c)
(V3)(V2),
cuts to defi.ne
(d)
^2/^3.
chapter 2
and Continuity
Functions, Limits
FUNCTIONS
A function is a rule which establishes a correspondence between two sets. For our
present purposes we consider sets of real numbers. If to each value which a variable x
can assume there corresponds one or more values of a variable y, we call y a function of x
symbolizing the function while
and write y - f{x), y = G{x), ... the letters f,G,
=
a.
function
at
x
value
the
G{a),
denote
the
of
f{a),
.
The set of values which x can assume is called the domain of definition or simply
domain of the function; x is called the independent variable and y the dependent variable.
If only one value of y corresponds to each value of x in the domain of definition, the
function is called single-valued. If more than one value of y corresponds to some values
Since a multiple-valued function can be
considered as a collection of single-valued functions, we shall assume functions to be
single-valued unless otherwise indicated.
of X, the function
Examples:
1.
is
called multiple-valued.
If to each number in 1 S a;
correspondence between x and
^
ic"
we
associate a
number y given by
defines a function /
which
is
x^,
we can
then the
single-valued.
f(x)
x^.
If
Note that although a function is often defined by means of a formula as in Examples 1 and 3, it does not have to be, as seen in Example 2.
For convenience we shall often speak of the function f{x) rather than the function /
whose value at x is f{x). The distinction should however be kept in mind.
GRAPH
of a
FUNCTION
of a function defined hy y = f{x) is a pictorial representation of the funcand can be obtained by locating on a rectangular coordinate system the points defined
The graph
tion
{x,y) or
[x,f(x)].
BOUNDED FUNCTIONS
If there is a constant
numbers),
we say
M such
that f(x) is
upper bound of the function.
^M
20
CHAP.
2]
21
If a constant m exists such that f{x) g m for all x in an interval, we say that
f{x)
bounded below in the interval and call m a lotver bound.
If mS/(a;)gM in an interval, we call f{x) bounded.
Frequently, when we wish to
indicate that a function is bounded we shall write |/(a;)| < P.
is
Examples:
= 3 + a;
greater than
f(x)
1.
is
4).
bounded
-1
in
lower bound
a;
1.
An upper bound
2 (or any
is
number
less
4 (or
is
than
any number
2).
2.
f(x)
it
If f{x) has an upper bound it has a least upper bound (l.u.b.); if it has a lower
bound
has a greatest lower bound (g.l.b.). (See Chapter 1 for these definitions.)
MONOTONIC FUNCTIONS
A function is called monotonic
X2 in the interval such that
strictly increasing.
Similarly
f{xi)
> f{x2),
if f{xi)
increasing in an interval
Xi<X2, f(xi)^f{x2).
^ f{x2) whenever
it is strictly
Xi
If
if
fixi)<f{x2)
f{x) is
if
decreasing.
2/
is
x,
The function
2/
MAXIMA
and
Example:
a;
sin
in
a;
-1 S
MINIMA
deleted
borhood of
a relative
TYPES
1.
xo
where
maximum
of
>
[i.e.
for
(or relative
all
x such that
minimum)
0<\x-Xo\<
8],
then f{x)
is
neigh-
said to have
at Xo.
FUNCTIONS
where
ao,
polynomial
if
aox"
a are constants
aix"-!
and w
is
...
Un-ix
(i)
tjie
degree of the
aov^O.
[CHAP.
22
2.
where
po{x),
Pl(a;)r~'
=
+
fix)
Pn-lix)y
form
Pn{x)
(2)
are polynomials in x.
.,pn{x)
3.
i.e.
do not satisfy
(2).
to integers, rational
Exponential function:
2.
Logarithmic function:
= a^
/(ic)
3.
see
3.
av^O,l.
Page
3.
Trigonometric functions:
sin X
smx,
cosa;, tana;
= ^^^,
^^^
cos
_ 1 _
" teE " '^IH:^
radians = 180). For
^Iir^'
a;
^^ ^
"^
,
^^^=^-^3^'
(b)
(c)
5.
=
=
=
sin-^ic,
^y^
g g
{-it/2
cos-i X, (0
1/
tan-la:, {-kI2
7:/2)
tt)
<y
<itI2)
(d)
(e)
(/)
i/
=
=
=
cot^
sin {-x)
cos {-x)
tan (-)
The following
Inverse trigonometric functions.
nometric functions and their principal values.
(a)
sin^
csc-^a;
sec"!
a;
cof^a;
= csc^ x
= - sin
- cos x
= - tan
a;
a;
list
=
=
=
sin-il/a;,
is
cos"'
7r/2
real
^ y ^ 7r/2)
(0 g ?/ g
(0 < 2/ <
{-7t/2
tt)
1/a;,
tan-^a;,
tt)
sinha;
-^
e^
(d)
cscha;
(6)
coshx
,, , ,
tanha;
(c)
^^^
1
= ^^^i^
e^-e-"
sinhx
= ^^^j^ = ^.^^-x
e'
P
^^r+i^
(e)
secha;
= ^^^^ =
,.,
x
e^ + e'
cosh
__
_
- -^^^^ - g._g-x
(/)
^
cotha:
,,
a;
csch^a;
CHAP.
FUNCTIONS, LIMITS
2]
sinh
cosh
{xy) =
{xy) =
cosh y
sinh x cosh y
cosh
a;
AND CONTINUITY
cosh x sinh y
sinh a; sinh y
sinh {-x)
cosh
6.
23
(-a;)
tanh(-a;)
= - sinh
= cosh x
= -tanh
a;
a;
tanh X tanh ?/
(a)
cosh-
(b)
a;
In
(a;
+ V' + 1),
alia;
In
(a;
a;
\/^^^),
csch-^a;
in
sech-^
In
(d)
(e)
a;
^A + V^^ll\
^
+ VI
a^M
q
tanh-^a;
(c)
LIMITS
of
= ^in(i|j,
|a;|
<
^in^^Y
coth-^a;
(/)
^j
|a;l
>
^o
<
3.
FUNCTIONS
we can
(however small)
<
find
some
number
positive
(usually depending on
.)
such that
exclttde x
= Xo).
Example:
Let
/()
if
We
closer to 4.
a;
= 2-
'^^^^ ^^
^*
f(x) = /(2),
1-4)
i.e.
'='^''
* ^
To prove
4.
is satisfied.
For
definition.
The
this
we must
by
x approaches
(i.e.
a;
limit
is
would
2), f(x)
gets
see whether
Problem
10.
f(x)
When
RIGHT
it is
unique,
i.e. it is
and
If
a;
-+ xq
and
left
hand
The
lim f{x)
x -+ Xo
h,
we
call
We
limits of f{x) at Xo
have lim
^-*^o
/(x)
if
and only
if
lim /(x)
x-tx+
lim fix)
x-*x-
I.
24
THEOREMS
LIMITS
A and lim
[CHAP.
on
If lim f{x)
g{x)
B, then
X-*X,j
X-fXf,
+ g{x)) =
1.
lim (f{x)
2.
\im(f{x)-g{x)) = lim
3.
limf/(x)g(a;)^
4.
lim^
x^x
lira f{x)
- r"
lim
fl'(a;)
lim g{x)
= A+B
lim
f lim /(a;)Ylim
lim /()
f(-.
/(a;)
flrC*)
=5B4
3^(0;)
flr(a;))
A-B
AS
IfB^O
X^*Xq
INFINITY
sometimes happens that as a; ^ Xo, f{x) increases or decreases without bound. In
such case it is customary to write lim f{x) = +00 or lim f{x) = - respectively. The
It
X-+-XO
X-*Xfl
symbols +=
written
(also
respectively, but
it
In precise language,
=0)
we say
0<\x~Xo\<B.
lim f{x)
that
00
if
number
we
whenever
in general) such that f{x) >
(depending on
we
that lim f(x) = -00 if for each positive number
we say
Similarly
infinity
<
-M
whenever
<
|a;
- Xo] <
8.
Analogous
We
lim f{x)
say that
I,
or f{x)->
as
x^
+<x>,
if
number
we
X-* +00
x>N. A
i/(a;)
l\<e. whenever
SPECIAL LIMITS
sin
hm^
X
a;
,.
1.
,.
hm
1,
lim(l+-)
Xj
X-fM \
3.
hm
X-+0
e,
lim
cosa;
X-+0
X-+-0
2.
[1
xf"'
X-+0 +
1,
lim
x-*i
mX
CONTINUITY
Let
a;
lim f(x)
f{x)
== a;o (i.e.
in a 6 neighborhood of xo).
X-*Xo
f{x)
be continuous at x
xo.
^^-2J
3.
exist,
exist.
i.e.
/(.-To)
Equivalently
must
must
2. f{xo)
Jim f{x)
lim f{x)
1.
25
if fix)
continuous at
is
we can write
^,
/(lim X).
Examples:
1.
If
/(x)
= {/ ^^g
Hence lim
2.
m=
If
/(ck)
o;^
/(2)
^^'^'^
fails to
/(:)
/(2)
is
and
f{x) is
/(a.)
4.
But
form
/(2)
= 0.
= 2.
continuous at
.iH to
f"^"*^/*^^^/(^)
said
be discontinuous at these points.
/(a;)
c.
- 2.
and
^^
fix) is
c:Sufty ^r dTsco^ruitr"
^^''''
continuity,
"^
a:-/!fT"nv*Vn'71'^"^"^'''^'"**^^*
t^."^"""
"^
Not^tWt^-
t?at
" '^
.?L
RIGHT
"^
'"^^
Similarly, /(x)
Definitions
is
THEOREMS
Theorem
= .o
if
^Hm
in
on
1.
ft^tSui^^in
an
/(.)
= x, IPlim
terms of
CONTINUITY
restriction
---o-
fix)
/(.),
'
i.e.'^if
/(xo),
''
'^
i.e.
In such case
f^xJ^^fTl
fixo-)
'^^^
>
fix.)
'^^'>'-
INTERVAL
CONTINUITY
If fix)
llnttir
polynom-r"')- sif^anTcr.!
\ZV
is
^"^
IS continuous at x = x..
tion of a continuous function is
il'
*^^ ^^^^
ST^/a^
continuous at
-^^
^^
^"'""^"^
also
^(-)--
'"-^^ ->ts
hold
^^^
^"
= Xo and
^ = ^(2/) is continuous ^t y =
y,
a /^ncizo^ 0/ a /t.^c^io.1 or comvosite
This is sometimes briefly stated
u 'isas: ^ conimuous funccontinuous.
and
andift'-Tf
If i/o--/(a^o), then
th^ the function
function,
"^**^ ^-^(^-"^
call /(.)
we can
and
we
definition of
=^
..
/(x)], called
AconZuZfut^
26
Theorem
Theorem
4.
5.
interval about x
= Xo
in vi^hich f{x)
>
a;
= a;o
[or f{x)
and f{Xo)>0
[CHAP. 2
bounded
[or /(a;o)
in the interval.
< 0].
Theorem
6.
If a function f{x) is
[a, b] and if f{a) = A and f{b) = B, then correbetween A and B there exists at least one number c in [a, b]
is sometimes called the intermediate value theorem.
Theorem 8. If f{x) is continuous in [a, b] and if /(a) and f{b) have opposite signs,
where a<c<b. This is related to
there is at least one number c for which /(c) =
Theorem
7.
If f{x) is continuous in
Theorem 7.
Theorem
9.
If f{x) is continuous in a closed interval, then f{x) has a maximum
for at least
for at least one value of x in the interval and a minimum value
value
and
one value of x in the interval. Furthermore, f{x) assumes all values between
for one or more values of x in the interval.
are respecand
Theorem 10. If f{x) is continuous in a closed interval and if
there
exists
of
bound
(g.l.b.)
lower
f(x),
greatest
and
(l.u.b.)
upper
bound
least
tively the
related
to
=
This
is
m.
or
which
for
f(x)
interval
f{x)
the
of
x
in
value
at least one
-M
Theorem
9.
SECTIONAL CONTINUITY
a^x^b
a^x^b
and
X4.
Fig. 2-1
UNIFORM CONTINUITY
be continuous in an interval. Then by definition at each point Xo of the
interval and for any >0, we can find S>0 (which will in general depend on both e and
the particular point Xo) such that \f{x) - f{xo)\ < e whenever \x - xo\ < 8. If we can find
not on
8 for each e which holds for all points of the interval (i.e. if 8 depends only on e and
interval.
the
in
continuous
is
uniformly
that
say
f{x)
Xo), we
we can find
Alternatively, f{x) is uniformly continuous in an interval if for any >
Let
8>0
f(x)
such that
\f{xi)~fix2)\
<
whenever
|a;i-a;2|
<
where
Xi
and
X2 are
any two
Theorem.
the interval.
uniformly continuous in
CHAP.
21
27
Solved Problems
FUNCTIONS
Let f{x)
1.
(a)
for 2
a;
8.
Find
(a)
- 2)(8 - 6) = 4 2 = 8
not defined since f{x) is defined only for 2
The set of all x such that 2 g a; s 8.
/(6)
/(-I)
(6)
/(6)
a;
(e)
What
th^ dn
is
8.
where
t is
i.e.
(d)
(b)
(6
is
/(I
(c)
and /(-I)
The following
such that
2t S
8,
9.
8
8
f{^)
2.5
7.5
2.75
2.75
(2.5jit),''aT2.?5)''^'
points
* ig.
of a parabola.
for
''
of2(ldt7nf?r.*^
and that of f{x) m
o
fl^(a;)
swer
swelTart"s
parts llVand"
(6) and
2<:.<8.
Problem
1.
(a)
(6)
What
'of
de/nitn?
(c)
^'^
i'T
^
1.
(a)
(6)
The
l.u.b.
(c)
The
l.u.b.
of g{x)
is 9.
The
g.l.b.
of g{z)
is 0.
nM
'^^
j
,=***
"* """""''
"
'"^'^^
^^'"^
^^
(d)
Let
if =^ is a rational number
1 0, if X is an irrational number
Construct a graph of /(a;) and explain
why it
= 1 since f is a rational number
/(|)
/(-5)
- 1 since -5 is a rational number
/(1^1423) = 1
since 1.41423 is a rational number
=
/(V2)
since \/2 is an irrational number
ffa;^
^
' ^
(6)
()
(6)
= / 1'
^"^
"
is
^'"^
^(*)'
misleading by
fM
28
4.
[CHAP.
Referring to Problem
1,
(a)
^,
(6)
+ lQ + y =
We
B
1
Vl00-4(16 + y) ^
10
f-'{y)
2h
0.
^ ^/g^_
Fig. 2-4
Then, y
/-'()
V9 -
for
In the graph, AP represents i/ = 5 -I- Vs-^, BP represents 2/ = 5 - V9 - a;. Thus
S a; < 9, /"'(a;) is double-valued. This is seen graphically from the fact
each value of a; in
that every vertical line to the left of P and the right of AB meets the graph in two points.
The functions 5 + yj'i-x and 5 - yj^-x represent the two branches oi i'\x). The point
where the two branches meet (or have the same value) is sometimes called a branch point, in this
case at a; = 9, y 5.
(a)
g a; g 9. (b) Is
decreasing in
Does g{x) have a single-valued inverse?
+ ^9-x
is strictly
(c)
+ V9-X1 >
If
(see
y
a;
<
x-l.
If
- xi >
xi
<
is
strictly decreasing.
X2
then
- xi,
- 2)(8 - y)
then 2/ - 5 = V'9 - a; or squaring, x = -16 + lOy-y^
5 +
a single-valued function of y, i.e. the inverse function is smgle-valued.
single-valued inverse
In general, any strictly decreasing (or increasing) function has a
=
V9^^
is
Theorem
The
6,
Page
26).
the figure of
results of this problem can be interpreted graphically using
xi
mono-
and
6.
wh enever
it
(a)
f{x)
^^^^
'
a;
'
^^^
^^^^
Problem
'-*'
4.
x.
Fig. 2-5
(6)
shown in Fig. 2-6 above. If 1 ^ a; < 2, then [a;] = 1. Thus [1.8] - 1, [V2] - 1,
However, [2] = 2. Similarly for 2 g a; < 3, [a;] = 2, etc. Thus there are jumps at
The function is sometimes called the staircase function or step function.
1.
the integers.
Fig. 2-6
is
CHAP.
2]
29
7.
= /W =
Pig. 2-7
If
(6)
/"'()
tan-
tan I
Fig. 2-8
2/-/(a;)-taniB, then
a;
a^CrTr^r2-rabovr
= /-%) =
^ ^^^^
tan"'
^" ^'^^
tan"'
a;
is
obtained
oriente/rurat
(c)
(rf)
-W2
We
"
valued'^nverse.*
8.
*"'
'""^ "'
*""" "
''""'^'''
'^
(e)
Show
Tf
Z^-^S^\
prinSpaTvalue
= -^1,
- V^+1
*^" (x
that fix)
^
1m^
~1,
is
-W2
+ l)2/-l = V^
or squaring,
it is
Srv21trnolThrouottnT;r"T
9.
If fix) = coshaj
inverse function,
H^=
yVy^-i.
Since
Thus
i/
i(^+e-^),
a;
Choosingthe
(a;
Va;
e^'-ZyC+l =
In (i/ Vl/' - 1).
- Vy^~l =
a;
1).
{y
0.
(x
"
is
^-
+ 1)V -
it
2(a;
'^^'^^ ^(*>
is
l)j/
a;
-l^^'^-- f--on.
e'
2j/
V4y' -
yfy
ln(2/
i.e.
^^^
not the quotient of two polynomials,
r^^^^T^^^^
so that it
flowever,
and W2,
we can
V2/'-l)
or
cosh-^/
= in
(j/
also write
y^^^Ti)
The choice
we have cosh-
a.
LIMITS
10. If
^"^
ia)fix)
= x\
(&)/(^)={o;
^;|,
Z^-^^V:^ZV<\T-T<\':'
"'
''^^'^'
/(a:)
^^'^'-^"^^
4.
- ^
----)
-h
that
30
Choose
Then
|a;
|a;
as 1 or e/5, whichever
and the required result is proved.
Take
that
a;'
if
11. Prove
a;
make
to
4| <
\x^
we can
6,
choose
=1
|a;-l|
<
>
we can
find S
S.
>0.
4a;''
and
2x*
.
that
such .^
,
6a;'
<
whenever
- 2| <
jx
3a; 3
a;^
2a;^
6a;
+ a:' + 3
on cancelling the
^~
^^^^
+3 _
""
(2a:'
_.|
'
'I
'^T^
common factor a; 1 7^
Then we must show that for any e > 0, we can find 8 >
< a; < 2, a; 7^ 1.
< - 1| < 8. Choosing 8 S 1, we have
2x^
- ^\ <
merely coincidental.
is
<
56.
\x^
2x*-6x^ + X^ + S
a
= -o.
7^^
lim
"''^
when
<
= 2.
that
We
Then we have
# 2.
no difference between the proof for this case and the proof
is
exclude
.,
=2
a;
we wish
If
There
a;
2i
-4 <
we
3,
we can
.05,
to hold at
(6)
smaller.
<
S|a;
of interest to consider
It is
- 4| <
|x^
is
a;
[CHAP.
0.
such that
|2a;'
4a;''
- 3a; + 5| <
when
|a;
4a;''
C \x-S\
12.
Let
=\x-3
f{x)
For
For
12a;''
- 2a; - 5| <
3,
a;
3,
Graph the
function.
Find lim
(5)
-3) _ -1
a;
>
3;
= 1,
j/
a;
(c)
f{x).
Find
right, f(x) -^ 1,
lim f{x)
i.e.
= 1,
(3,0)
l/(a!)
- 1] <
To prove
this
we
we can
>
- 1 < 8.
-1
11
find 8
<
whenever
f{x) = 1 and
whenever
the proof
so
1 <
a;
F>S-2-9
As x^
Since
lim f(x)
from the
=
left,
f{x)^ -1,
lim
<
<
Making
lim_ f(x)
= -l. A
(fe).
0.
|a;-0l
i.e.
f(x),
If
<
<3
0<*-l<8.
We
5)
= l,
<
(3, 0).
from the
such that
13.
l2a;I
/()
a;-3
(d)
-3 ~
-(a;
\x--31
<
a;
X--3
(c)
6(l2a;''l
a;-3
la;--31
>
a;
a;
<
f{x).
As
- 2a; - 51
(b)
|2x''
x-*3
x-*3(a)
- 11
(a)
Find lim
(d)
fix),
|a;
^^
'
(0,
lim
- 3a; + 5|
Taking
>
0,
we can
find
>
such that
la;
sin 1/a;
0|
<
when
6.
la;|
<
8,
then
the choice
|a;
sin
= e, we
l/a;l
|a;l
see that
S
sin l/a;| <
|sin l/a;|
la;
la;l
<
when
since
<
|sin l/a;|
la;l
<
8,
for all
a;
7^ 0.
CHAP.
2]
Evaluate
14.
lim
To prove
X-H.O+
e-i/^'
we must show
this conjecture
2
1
Now
-^ > i
^
- 1)' ^
In (2/e
any value of
15.
>
>
0,
will
work
when
<
^'^
e-'''
(t" V'
*^a" ^^
"^
a;-
-^ <
16.
<
^ in(2/e-l) assuming
for
1)4
>
when
< e < 2.
>
=
^^^^
1)*
^^'^ ^"'^
2.
If
2,
0.
x^i{x~
^^
<e<
if
co
^ ^^^^^^
and prove
^""
P^^^^e the
-->-
(^^^P-d-
0<|a;-l|<s
<
when
all
lim
^^^ > M
such that
is
("
Choosmg
>
<
a;
e'/'+l
^'"'^
*^* our
find S
2-2-2e-'''
we can
^ ^
^'
Choosing
<
>
g"^
+ e-^"
that, given
+ ->/' -2I =
when
31
(x
- 1)* <
<
or
|a:
- 1| < J-
lim
^^
e-o
OD =
and
AC
Area of triangle
OAC <
Area of
f smi?cos# < ^e < Atanij
Dividing by 1 sin
OAO <
sector
as in Fie- 2 in hlw
ru
'""^
are perpendicular S, Oi.
Area
of triangle
OBD
e,
<
cos 9
^H*
<
cos
<
sin e
COS*
tan
<
cos
As
0,
COS e
and
it
^^ =
Fig. 2-10
THEOREMS
^^' ^^
il^o
We
By
on LIMITS
^^^^
^'^*^'
P'^"'^^
if
*^at
Jim
/(x)
>
it
must be unique.
h and
we can
\f{i>)~li\
\f(x)~h\
<
<
e/2
e/2
lim /(x)
find
>
when
when
k,
then
Z,
i,.
such that
<
<
<
\x
xo\ <
|a;
a;o|
32
Then by inequality
\h-h\
i.e.
18. If
\li-h\
lim
= B
lim g(x)
Since
\g{x)\
>
we can
find 8
g(x)
i.e.
19.
<
\B\
\\B\
Given lim
= A and
f{x)
(5)
lim /(i,(.)
(a)
We
= AB,
Using inequality
By
>
{A
>
\[f{x)
where
(6)
We
(A
have
lim fix)
Since
A - 1<
f(x)
a;ol
<
<
(c)
We
la;
ja;
a;o|
8.
\g(x)\
[f{x)
(^)
lim
+ g{x)] = A + B,
if
5^0.
such that
Q
<\x-Xo\ <
&
>
and
e/2
82
<
<
when
when
e/2
lg{x)-B]\
>
\g{.x)-B\
{1)
such that
<
<
|a;-a;o|
|a;
\f(x)-A\
a;o|
8^
(2)
8^
<
i5)|
81
=
^
S
AB\
e/2
and
- B] +
\f(x)\\g{x)-B\ +
\f{x)\ \g(x) -B\ +
f(x) [g{x)
8,
such that
A, given
>
<
when
|a5
aJo]
<
8^.
e/2
i.e.
we can
B[f{x)
- A]
lB||/(x)-A|
(\B\
1) \f{x)
l/(a;)-A|
<
~A\
<
for
< P where P is
such that
8^ >
|/(a;)l
find
>
we can
find S3
(i)
\
|a;-a;ol
>
<
\g(x)-B\
\f(x)-A\
such that
<
Sj,
i.e.
a positive constant.
8^.
lim /(x)
<
ia.
(S),
<
e/2P
g/i^i
+ i)
for
*^
83.
we have
in (i),
- xol <
- AB\
where
g(x)
P-^
m+
the smaller of
>
b\
we can
find 6
lf(f)~-^l
<
1)
81, 8^, 8,
>
2(|g|
"=
>
\g{x)
we can
B\
find 81
< \B\
is
complete.
such that
when
>
such that
when
1)
0<la;-a;l<S
\B\\g(x)\
e
hypothesis, given
<
is
J:__i| =
By
find 8i
A, we can find
\f(x) g{x)
<
B, given
Using these
for
<
for
lim
when
<\x-xo\ <
(a)
>
find 8
lim g(x)
<
^\B\
\\B\
5^0,
if
Since
- B\ <
<
B, prove
< A + 1,
Since
la;
g(x)
Zi
:^\B\.
\[f(x)-A]
8 is
|/(a;)
<
and
(1), (2)
we can
|/(a;)-A|
g(x)]
Thus
we have
\g(x)-B\
Then from
\g{x)\
we can
3,
B)\
hypothesis, given
<
|a;-a;ol
\g{x)
\g{x)\
- (A+B)| <
Page
2,
+
>
\B-g{x)\
\[f{x)+g(x)]
+ g{x)\
such that
iL^ = i
\[f(x)
>
e/2
zero.
such that
<
for
lim g{x)
(c)
>
e/2
must be
i|B|
from which
\g(x)\,
<
\f{x)-h\
we have
g(x),
IB]
\h-f{x)\
B,
B = B-
Writing
+ f{x)-h\
\h-f{x)
- 0,
3,
is less
^(a;)
Page
2,
[CHAP.
<
[a;
a;o|
<
Sj
(5)
CHAP.
2]
By Problem
18, since
lim g{x)
>
\9{<e)\
Then
if S is
Prom
parts
and
8j,
^ |gW-g| <
\B\ \g{x)
S^
and
(6)
we
0,
can find
<
when
l\B\
the smaller of
= B
we can
|a;
>
6,
a;o|
<
33
such that
8,
write
<
whenever
il^l
<
a;o|
proved.
is
(c),
lim
/(a;)
ff(a;)
= i^(=)
^"''
=
j t;^
g(x)
^'i = I
x'.i;x '
i'Ji;"|j
A\
\f(x)
la;
'"* * ^"
iS-r .T ^T!
foVtt^ror/sS^rcVind^x'
"^'^^^ '"""^^^^
^^
^"^
and
,/2
l/(*)
"'^"^
<
a;^ , a;-^-=o.
\g(x)
- b'\ <
e/2
so that
^^^"^^
"'
^-
^ ^^"-
f^t
^"^^^
20.
(a)
lim (x^
- 6a; + 4) =
lim
a;^
(2)(2)
lim (-6a;)
(-6)(2)
lim 4
= -4
(c)
lim
i--i
a;''
-oo
+ 3a;
6a;''
+ 3)
lim
- Sa;" + 1
+ a;' 3a;
2a;*
lim
(a;
(^|H2Z_1)
CO
(a;='
lim
\/4
fe
lim 2
lim 6
lim
sill u^
Note that
-I +
lim
- +
X
2_
lim
lim
-|
i-n
a;"
lim
lim
V4 + fe +
A/m+2
^
"-">
lim
(-3)
-4
a;'
lim
x-oo
V4 + ft -
=
1.
19.
()
6+^-4
-K
(d)
X*
a;
by Problem
- 1)
(2a;
+ 3a;-2)
a;*
lim
*-+
lim
4 + A-4
A(V4+1 + 2)
sin
a;
'
v^ =
CONTINUITY
21.
Prove that
Method
f{x)
ia;
x" is continuous at
a;
= 2.
1:
By Problem
Method
lim
10,
/(a;)
/(2)
and so
f(x)
is
continuous at
a;
= 2.
2:
|/(a;)
v
i
'
-/(2)|
m ;i
[CHAP.
34
22. (a)
Prove that
From Problem
continuous at
is
lim f(x)
13,
.;
= 0.
= 5,
so that f(x)
discon-
is
23. Is the
function
lim f(x}
f{x)
= -8
Problem
(see
Prove that
if
fix)
(c)|||
{b)f{x)g(x),
if
not continuous at x
is
becomes continuous at x
11), it
= !?
continuous at
^^zii
/(I) does
a;
a;
= Xo,
so also
are
<\x-xo\<
and rewriting
Prove that
We
/()
Prove that
f{x)
Since X
is
cb,
finally 2x'
By
8.
is
a;
<
as |aj-Xo|
8,
including x
i.e.
choosing
2x^
+ g{x),
+x
is
=e
A = f{xo)
and
B = g(xo)
x<,.
that, given
must show
f{x)
(a)
/(xo)^O.
taking
These results follow at once from the proofs given in Problem 19 by
26.
re-
By
(b)
25.
Can one
(6)
= 0?
But
0.
a;
not continuous at
is
^J
0.
tinuous at x
24.
|^^i"l/^'
f{x)
]/()
-/{o)|
|a;
<
a;o|
a;
= Xa.
continuous.
27.
Prove that
If X, is
lim
= ^x^^
f{x)
if
V^"^^ =
and
/(5)
for
<
a;o
<
9,
then
28.
xa.
An
e,
8 proof, directly
a;
lim
from the
in the
9,
then f{x)
lim f{x)
Hm yf^^ =
/(9)
yfW) =
definition,
is
lim
y/x-h = Vo-5 =
Thus the
Hm
/(^)
fi.x,).
Also,
result follows.
= V7(^
if
/(^)
is
continuous
domain of
continuous.
(a)
f(x)
fM ^
rW
(c)
fix)
/6^
i^
l
+
+
<="'^
sin
^ns.
all
x except
Ans.
all
Ans.
all
> -10
Ans.
all
a;
#3
Ans.
all ^,
x-l
is zero)
a;
J
= Tr-
VlO +
l0-i''<^=^~3'^
a;
(d)
f(x)
(e)
/W =
(/)
/(a:)
If
10-''^""''"' *
J|
(see
since
Problem
55)
/(3)
Umf{x)
^r
a;
>
Then
0,
/(a;)
f{x) is
= ^^^ =
0.
continuous for
If
all
a;
< 0,
/(a;)
x except
a;
=
= 0.
^^ =2.
At
a;
= 0,
/(a;)
is
undefined.
CHAP.
2]
fi.x)
(ff)
""
a;==0
2,
As
in
f(x) is
(/),
continuous for
^^M
lim
follows that f{x)
it
Thus
is
= xc&cx = -^.
/()
a^csc,
/(O)
continuous for
all
continuous for
f{x) is
a;
< 0. Then
+ =
li^
f(x)
{h)
35
all
Ans.
a;
all
Since
1.
S 0,
lim
esc
a.
a;
li^^
[compare
/(O)
domain of
its
a.
= 0.
everywhere in
i.e.
x except
at
left)
lim 2
definition.
....
^,
^(o),
(h)].
UNIFORM CONTINUITY
29.
Prove that
Method
1,
f{x)
x^
is
uniformly continuous in
<
a;
<
1.
using definition.
We
where
If
lx'-xl\
la;
a;| (a;
<
a;o|
+ 1(
11
la;
Choosing
i^'-*l<2^r.''"a;f<'rwwAV"T*,"*
"^ " ' ^""^ ""* "
'
0<a;< J
The above can also be used to prove that f{x) = x'
"^
^''-
is
- xl\ <
\x'
depends only on
Prove that
Method
/(x)
1/a;
is
- a;|<
a;oI
|a;
'
Let
a;
and
Xo
However,
0</<r
^" ^'^^
" ''^^
uniformly continuous in
<
a;
= ~-^.
Then
1
J^
|a;-a;o|
i.
<
it
^(^>l
"
<
1.
-'^^
'^
= T >
Xi,
^ contradiction and
'^^^^
" -I <
<
(since
follows that
<
m^l/x
<
e>0 we
^^^ " -
"
should
S.
1).
2:
Let Xo and
a;o
f{xo
(0, 1).
Then
Xo
a;o
+S
'
xo\
H^'^^^
""**ly continuous
Ju^^Ta^.^^ ^^""l''*"""
i'
L tt interval. ''"
'
1:
bp
Method
\x
'
Method 2:
The function fix) = x' is continuous in the closed
interval OSa;^!.
Page 26 it is uniformly continuous in S a; g 1
and thus in < a; < 1.
30.
when
xo{xo
+ S)
--
*^
^-et-
[CHAP.
36
MISCELLANEOUS PROBLEMS
3L
If
y^f{x)
is
prove that
continuous at y
= yo where
yo
= f{xo),
and yo respectively,
Since by hypothesis f{x) and g(y) are continuous at xo
= g{f{x)}.
Let h{x)
we have
Then
lim f(x)
/(lim x)
f{xo)
lim g{v)
g(l\m y)
gM
lim hix)
32,
is
z = g{f{x)}
lim g{f(x)}
g{f(x)}
is
flr{
lim /(x) I
continuous at x
ffifM)
g{f(xo)}
/^-(sco)
= xo.
8, Page 26.
must be an interval (a a+h),
Suppose that /(a) < and f{b) > 0. Since f(x) is continuous there
and so has a least upper
bound
upper
an
has
a+h)
{a,
points
of
h>0 for which fix) < 0. The set
if /(c)
because
bound which we call c. Then /(c) S 0. Now we cannot have /(c) < 0,
T^^f/^^^*"^^ ^^^
c) for which f{x) < 0; but smce
than
greater
values
(including
c
about
would be able to find an interval
and so we must have /(c) - as required.
c is the least upper bound this is impossible,
Prove Theorem
>
If /(a)
33. (a)
Given
and
f(x)
< 0,
2a;^
- 3rr^ + 7x - 10,
X in
f(b)
(c)
Show how
Prove that
(b)
/(2).
/(a;)
(b).
(a)
/(I)
(b)
If
2(1)'
- 3(1)^ + 7(1) - 10 =
-4,
/(2)
if /(a)
2(2)
and
- 3(2)^ + 7(2) - 10 =
/(b)
have opposite
8.
(Problem 32).
continuous in
realize that the given polynomial is
only
need
we
theorem
To apply this
exists a
there
Thus
0.
and
>
/(I)
<
/(2)
that
(a)
iaa;g2, since we have already shown in
0.
/(c)
that
such
2
and
1
between
number c
f(l 5^
2(1 5)
- 3(1.5)^ + 7(1.5) - 10 =
lies between 1
has a value closer to
worth pursuing in practice).
= 2(1.4) - 3(1.4)^ + 7(1.4) - 10 = -0.592 we conThus we consider x = 1.4. Since /(1.4)
1.4.
is most likely closer to 1.5 than to
clude that there is a root between 1.4 and 1.5 which
to 2 decimal places.
Continuing in this manner, we find that the root is 1.46
= 0.5
/(1.5)
34.
Prove Theorem
10,
of
Theorem
hypothesis,
f{x)
=M
i
4,
Page
Page
we can
> i
26.
26.
find
x such that
so that
However
we must have
.,
if
^,,
M - f(x)
is
we suppose
<
by
M.
also continuous.
M - f{x)
is
continuous, by
we must have
J^^J,^)
Similarly
we can show
that
that there exists an x in the interval such
f{x)=m (Problem
93).
^^^^2]
37
Supplementary Problems
FUNCTIONS
Give the domain of definition for which each of the
following functions
35.
Ans.
If
36.
(6)
all
{a)
-2Sa;g3,
x>3,~2<x<2.
find:
If
fix)
iQg
/A
(/)
^^0;
<
s 2,
find
2^/(0)
h^O;
f(
5X
^^^
a;
Determine whether
^^^
^^^_ ^^^
(.)^.,^o,2
|,.^0,2
(d)
(a)
the
l.u.b.
f(x)
(6)
8,
m=
(a)
(b)
-3 S . ^ 3
1x1.
(/)
= 2-i^, -2Sa;S2
fix)
/(x)
(fif)
CO,
where
(e)
/(x)
a;2
sin
[.]
greatest integer
coshx
x<0
sin
39.
^^^_^^^.^
/{/(a:)}.
(/)
|^,,^04,2
and single-valued:
(d)
log. (a,' _ 3a, - 4a: + 12).
s (2m + 1W3, m = 0, 1 2 ',
'-^-5,2
-^+5-.
= 2a;^
5 /(-I)
2m;./3
(c)
is real
V^i^,
(c)
2,
M^m
(,)
_^
7^
a;
(.)
(e)
^-Wi
38.
(x
(d)
/W + /(4/.;),
- 2)/(x^ - 4),
(6)
/(.)=f^,.^2,
(d)
37.
+ 4),
V(3-:)(2x
(a)
1/a;,
a;
.^
f(^x)
(^^
a;
^'"'^
where
40.
(a)
cos
From
r
,3
the choice
(6),
find
cos-
(1/2)
- cos- (-1/2).
41.
Work
42.
43.
44.
parts
(a)
and
(6)
of Problem 40 for
(a)
2/
6"^,
(6)
2/
(a)
In
|a;],
seC' x,
(c)
sin- (-V3/2)
(6)
(c)
(/)
tan-
(d)
cosh-y2
(e)
e-e..- <.v.,
(^)
(fe)
(i)
(^.)
Ans.
45.
2/
(c)
-^/3
t/2
id)
ln(l
(a) cos {^
(a)
-:./3
(&)
Evaluate
+ V2)
(6)
whv
U^^^
(6)
2/
e'l-l sin
cot"' .
b),
a;.
23, evaluate:
^i^-: ^
sinsin-
-1 g g 1
S S ^/2
W2 g g 3;./2
^^^-,
(cos 2a;),
(cos 2x),
a;,
a;
a;
tanh(csch-3x), a;#0
eos(2tan-a;^)
- 2a;
(e)
(^)
^/2
(/)
^/2
ih)
2a;-3W2
cosh"' {coth
ExplZ
(In 3)}.
Ans.
i_^4
\x\
^'^
(a)
W^^+t
-^2/2,
(6)
In 2
^^^
1+^
'
[CHAP.
38
tan-
Prove that
(a)
46.
tan-^a;
Is
47.
If
f{x)
48.
4<J
Prove the
tanh ^x
(sinha;)/(l
cot'' b
coshx),
= 1.
/(O)
(sin^x)/a;S
a;^/
= 1.
cot"' a.
3 sin x
(e)
In esc
|
4 sin^ x,
a;
(a)
of:
4 cos'
a=
3 cos x,
In |tan ^x\.
undefined
is
cos 3x
(c)
cota;|
50.
(^^^,
sin 3x
(b)
sech^ x.
Ry)
identities:
if
also?
f(.x)
tan"' 6
are taken.
the conventional principal values on Page 22
Explain.
= W2
:.
'^/2
ST- tanh^=^
(d)
cot-
tan-'(l/a;)
(6)
:.
fix)
or
m^l;
.(sinx)/a;
defined but
/(O) is
f(x)
(6)
1.
LIMITS
by using the
first
limits,
51
(a)lim(x=-3a.
(/)
li^.
Let
f{x)
Evaluate
- 1,
0,
lim
'"'
(6)
.,
+ 5,
a;
Ans.
/(fe)
- /(0+)
{c)
lim^^zT^'
4,
(d)
-1,
(0
(a)
Construct a graph of
(d)
/W,
id) 5,
(e)
_d
ana
/feN
(O)
54.
(a) 2,
(a) If
Mm
fix),
-1,
if)
Prove that
5e.
Let
we
if
/(.)=|lp,
Find
lim 1^,
(a)
_o+
(6)
a;
59.
^f^ =
60.
Prove that
61.
Explain
62
If
If
[.]
64
If
lim fix)
What
m=
.#0,/(0)
(f)
Urn
fix).
is
y.(^)
j^
is
lim !^
Evaluate
(a)
-1,
(c)
(a) |,
(6)
Illustrate
lim sin
(a)
Hm
(6)
lim 10 -'
(&) 1/6,
(c)
2,
a;
(1
= A
and
(a)
u2{fix)-9ix))
Km /W.
^Um
(c)
/(.),
lim
justi-
/(.),
lim/(la;|) exist?
-,
:=
lim
(a)
(e)
x.
^ljm^^ =
/(.).
Ans.
(a) 1,
(6)
-1
lim gix)
(6)
ib)
(a)
=T-B,
ib)
B,
lim 3^:r2
lim e-sina;
does exist.
Umjix). (0 Jn^
Urn {.-[.]},
Jim [fix)}^
Explain.
/(*),
id)
Um
fix),
(e)
Ans.
(a)
Hm
/(.).
evaluate
(a)
(c)
= A\
true?
generalizations of these do youluspect are
lim fix)
Explain.
A. prove that
If
= 2?
(a) still
writes:
- e-)
(b)
0,
(d) 1/6,
/(0)
(6)
a;
-,
:.#0,
cc^cosl/*,
= i.
Ans.
i-o-
largest integer
63
65
(a)
why
(a) 2,
16
/(,)=.3^+W,"Evaluate
Arts,
I'
'
limits.
h)*
/(a;) is
If
lim
all cases.
58.
(a)
consider
lim 10-i/c'-3>'
fying answers in
57
evaluate
x^ cosl/x,
^'^
f ih)-fiO-
li^
^nm_
(2
,.
/(a;).
fix),
lirn
ie)
(/)
(6)
fix)
32.
(e)
jj^
Atis.
\/S-2
^I^.-J^T
^'^
-1,
-10,
(c)
(6) 9,
lim
(a)
(6)
(0 MmJ(x),
/(.),
....
.
^
Evaluate
<
=
>
a;
in each case.
53.
(a) 2,
Arts,
r SiB
52.
WUm^^^,
+ 2),
definition
a,i_4
(6)
(6)
Jhn
Hm
VM
{.
[a.]}.
0,
^/A-
prove that
lim {afix)
b gix)}
= aA + bB where
a,b
any constants.
(6)
<^^^P-2]
66.
hm
(a)
(3i;-l)(2a;
(b)
^^'
lim
i
^'^
i!l?}^(^
(5a;-3)(4a;
If
2 -3a;
Given
lim
Ans.
(a) 3,
X-
io
If
lim
()
72.
limits.
0,
fe
(">
-8/21
(*)
3/10
1/32
Ans. 1/12
x').
lim
-x g(x)
evaluate:
('^^
- 3)
(=
i,
(c) 1/2,
(6) 0,
-+0
gX
= B ^
^1^^^=:^^
^"'iro-^^
^^^
Let 8
(Hint:
lim g(x)
"*'
1,
()il"3^
-.
^''^
and
^"*-
3^)
-^^
= A
/(a;)
"
= ABC.
+ 3)
+ 6)
""".^
k-fO
71.
68.
g^g'
lirr,
'
70.
and
(b)
67.
69.
A,
A+B + C,
39
^">1
CSC
(d) -1/^,,
4?^-^^
+ 3 sin
(e) lira
^*
(e)
(/)
,,,
(^)
40;
^os^a^-cosfcg
lim
(/)
2/7,
2a;
^^^
i(6-a),
(p)
-i,
- 2cos. +
i!
:-+o
cos2x
a;^
^.^ 3 sin ^x
sin
3:ra;
(h) 4^
Prove that
prove that:
1,
W l^ =
= *-;
/()
if
and only
lim
if
Inf
f{x)
>
a,6
0;
lim
(c)
lim /(x)
^^
a.
I.
CONTINUITY
73.
Prove that
/()
a;^
74.
Prove that
/(a;)
1/a;
75.
- 3a; + 2
is
(a) /(.)
^;
(6) f(x)
a;-la;|;a; =
4n.
,^0,
continuous at x
continuous
/(0)
= 0;
(a)
at
=,
= 4.
2,
in 1
(6)
(o)
f(x)
a;
(a)
discontinuous,
(6)
'"*'^^''
Prove that
78.
If
f{x)
= x'
m/g(x) and
is
80.
liiiio"^
"''
(tan"'
'"
(c)
continuous,
(<)
= 3;
a;
=2
^,
*"
l<a;<2'
a;
discontinuous
Prove that
''
continuous,
/(2)
[x]
1< < 2,
a;
79.
f(x)
x^2,
r-
^llZ H*'''*
77.
3.
g^;
/In
'"
(6)
is
a;)/a;,
is
/(O)
=1
a;
=
is
a;.,
continuous at
a;
a;
a;..
= 0.
/(a;)/^(a;)
is
a;
= a;.
for which
82.
functions.
Give the points of discontinuity of each of the following
(6) f(x)
An.
(a)
a;
# 0, /(O) =
a;
none,
2, 4,
(b)
UNIFORM CONTINUITY
Prove that /(sc) = z' is
83.
= x^
(c)
(d) /(a^)
none,
(d)
7,r/6
uniformly continuous in
a;
<
<
a;
<
is
(c)
any
finite interval.
a< <
a:
<
a;
<
"
if
aS 0,
(b)
uniformly con-
1.
and
(a) f{x) g{x)
interval, prove that
State and prove an analogous theorem for
If f(x)
(6)
0, 1, 2,
>.
85.
is
a;
{b)0^x^ 2,
2,
Prove that
f{x)
'
+ 2 sin
2m,r, m =
2m-, II77/6
<
(a)
84.
86
[CHAP.
40
f(x)g(x)
m/g(x).
MISCELLANEOUS PROBLEMS
87.
Give an
88.
(a)
"e,
<
3,r/2
a;
<
5ir/2,
5n-/2
<
a;
<
a;
a;
77r/2,
89.
90.
(a)
(6)
a;
is
a;
tan x
(c)
and
^
y^x
and
4.49 approximately
= 0.
=
has infinitely many real roots.
a; + 1
Prove that for large values of x the roots approximate those of cos
Prove
92.
Suppose
.,
that
h),
,.
lim
j:_o
f(x) is
a;''
sin
(l/a;)
^-^
smx
a;
0.
0.
continuous at
where h>0,
|/(a;)-/(a;o)|
93.
a;
91.
xo
r/2<x< 3ir/2,
in
< ^ f(x<,).
bound
(a) Prove Theorem 10, Page 26, for the greatest lower
Page 26, and explain its relationship to Theorem 10.
(see
Problem
34).
(6)
Prove Theorem
9,
chapter 3
Sequences
DEFINITION
of a
SEQUENCE
Examples:
1.
The
Un
2.
The
Un
f(n)
numbers 2,7,12,17,
.,32 is a finite sequence; the nth term
= 2 + 5(w - 1) = 5n - 3, n = 1, 2,
7.
set
of
set of
numbers
l/(2n-l), w
LIMIT
we
1, 1/3, 1/5,
1/7, ...
is
is
given by
an
sequence with
infinite
Jith
term
1,2, 3,
SEQUENCE
of a
if
\Un
I.
n~*oo
Example:
m = 3 + l/w
lim u = 3.
If
(3n
the sequence
l)/w,
is
4,7/2,10/3,... and
called divergent.
unique.
i.e.
if
a limit exists
it
is
A more intuitive but unrigorous way of expressing this concept of limit is to say
that a sequence Ui, Ui, Us,
has a limit I if the successive terms get "closer and closer"
to I. This is often used to provide a "guess" as to the value of the limit, after which the
definition is applied to see if the guess is really correct.
.
One should observe the similarities and differences between limits of functions and
sequences. In defining lim f{x) = I, the limit I is attained for all possible approaches to
infinity.
I,
the limit
n-+oo
THEOREMS
SEQUENCES
If
on LIMITS of
lim a = A and lim
1.
lim (a
lim a
2.
lim
lim a
+ b) =
(a &)
=
B,
lim b
then
lim 6
= A+B
= AB
41
SEQUENCES
42
3.
lim (&)
4.
lim y-
(lim
lim a
ttn
,.
n~*
=B
v^
A v^ 0,
and
A = 0,
lim
^-^
may
or
may
not exist.
A,
for p
= any
real
number
if
A"
p^,
for p
= any
real
number
if
p^ exists.
exists.
n-^oo
==2)"^""
lim p""
6.
6n
and
(lim an)"
^
n-+oo
i^
B=
If J5
lim a^
CO
If
5.
iTzrir
= AB
(lim b)
ff)
[CHAP.
INFINITY
We
ber
write
lim an
(depending on
<x>
M) such
number
if
that
a>M
for
all
n> N.
n-^oo
M
M
quence
If
is
m^Un = M
Un-n
= Un
is
is
the sequence
convergent sequence
If
is
the sequence
is
called
monotonic increasing;
if
u+i>Un
S P.
Every
it
is
called
if
u+i<Un
strictly increasing.
it is
1.
is called
is
It is also
strictly increasing.
2.
The sequence
1,
1,
1,
1,
1,
Is
ing.
3.
The sequence 1,
However,
it is
is
bounded above.
number
and
of a
SEQUENCE
is
if
Un
= M,
Un
= m,
if
CHAP.
SEQUENCES
3]
definition of I.u.b.
and
43
g.l.b.
numbers
for sets of
in general (see
5).
number I is called the liTfiit superior, greatest limit or upper limit (lim sup or lini)
of the sequence {m} if infinitely many terms of the sequence are greater than I - 1 while
only a finite number of terms are greater than l + t, where is any positive number.
number
called the limit inferior, least limit or lower limit (lim inf or lim)
is
only a finite
if
number
infinitely
positive
+t
while
number.
These correspond to least and greatest limiting points of general sets of numbers.
If infinitely many terms of {m} exceed any positive number M, we define lim sup {^t}
CO.
If infinitely many terms are less than ~M, where
is any positive number, we
If
If
lim Un
lim Un
= =, we
n-+
= <.
define
we
define
define
oo.
= <.
oo
is
it
always has a
finite
if
and only
if
lim sup
lim inf m
is finite.
NESTED INTERVALS
Consider a set of intervals [a, b], n = l,2,3, ..., where each interval is contained
and lim (a - bn) = 0. Such intervals are called nested intervals.
We can prove that to every set of nested intervals there corresponds one and only
one real number. This can be used to establish the Weierstrass-Bolzano theorem of Chap. 1.
(See Problems 22 and 23.)
CAUCHY'S CONVERGENCE CRITERION
Cauchy's convergence criterion states that a sequence {} converges if and only if
for each >
we can find a number
such that \Up -Ug\ < e for all p,q> N. This criterion has the advantage that one need not know the limit I in order to demonstrate
convergence.
INFINITE SERIES
Let
Si
where
= Ui,
S2
S, called the
The sequence
Ui
be a given sequence.
+ U-z,
Ss
= Ui+U^ + Ua,
Form a new
is
is
the
sum
.,
sequence
Si, S2,
Ui+Uz + Ua-i
of the first
n terms
-S3,
where
\-Un,
symbolized by
00
Ul
which
is called
an
infinite series.
+ U2 +
If
Us
lim Sn
= S
^ Un
n=l
exists, the series is called
convergent
to sequences is given
SEQUENCES
44
[CHAP. 3
Solved Problems
SEQUENCES
1.
Write the
first five
'*'**
^'\sn-{-2j
fl
( D"!
,
(&) i
0, gi, 0, gs
Ans. ^,
2' 2'4' 2'46' 2'4'6'8' 2
K-[
2>4-6--2nJ
(c) i
o.A. I
(^){f + 1 + 1 ++!;}
^^'
^'^^
(-!)"-' a;"-' 1
(2m-1)!
Ans. jj,
g
'^'*'
i+
i'
-jg"
1!'
-'
a;"
3!
5!'
'
4+i+
i'
7!
We
define
i+i+4+
4'
10
i + i + i + i^ +
i^'
'
9!
'
= l'2'3-4..-%. Thus
= 1.
Note that n!
etc.
= 1,
1!
3!
= 1'2'3 =
6,
5!
= l'2-3-4'5 =
120,
Two
students were asked to write an nth term for the sequence 1, 16, 81, 256,
and to write the 5th term of the sequence. One student gave the wth term as Un n*.
The other student, who did not recognize this simple law of formation, wrote Un =
lOn^ S5n^ + 50n 24. Which student gave the correct 5th term?
which agrees with the first
If u = n*, then
mi = 1* = 1, mj = 2* = 16, M3 = 3* = 81, M4 = 4* = 256
.
ttn
first
ui = 1,
then
U2 = 16,
Us = 81,
U4 256
= 5* = 625.
= 601.
Both students were correct. Merely giving a finite number of terms of a sequence does not define
a unique wth term. In fact an infinite number of mth terms is possible.
LIMIT
3.
SEQUENCE
of a
sequence has
its
3^
17?\
-i
(a)
Write the
100th,
1000th, 10,000th and 100,000th terms of the sequence in decimal form. Make a guess
as to the limit of this sequence as w-* >. {b) Using the definition of limit verify that
the guess in (a) is actually correct.
(a)
% =
.22222
M = 10
.64444
.56000
We
on
c)
Now
|m
3n 1
r-3
4m + 5
-;
I
-||
<
for
all
>
|4(4w
= J(19/4e
complete.
Choosing iV
is
Note that
if e
= .001
when
4.5>g,
5), we
see that
(for example),
+ 5)| <
i^>l,
and the proof
10,000
.74988
it is
n=
100,000
.74998
small) there
is
a number
(depending
n> N.
-19
3
-r
how
(no matter
Note that
n =
1000
.74881
(6)
n=
n-lOQ
.73827
term
|m
19
4(4m +
5)
>l@-5)
fl <e
for
N = 1(19000/4 - 5) =
differ
from f
all
n>N,
so that
lim m
''"*
=f
in absolute
CHAP.
4.
SEQUENCES
3]
lim =
Prove that
We
on
e),
we
^
Prove
see that
,,
We
n>N.
(5
+2-10"
10")
-3
>
1/c,
10"
(a)
is
an> M
for
all
>
>
10"
7/3e
a number
<
3-10")
n>N.
(depending
0.
+
+
such that
- 5,
when
>
10"
is
2'10''
S-IO"
<
for
all
>
we can
iq")
n >
or
0,
''^^^'^^'i
''
''
log.o {i(7/3e
i.e.
- 5)} =
N,
If
a 7/15, we
see
we can
we
write
a.
(1
- 2n) =
->.
(depending on
M) such
that
00,
(&)
lim
number
M,
n > if^^^^
+ l)
' \ log 3
/
00
00
- 1)
(2n
lim S^"-!
(a)
find a positive
lim
write
when
a<M
n> 0.
all
we
then
result.
7/3e - 5 >
real only if
4-
3(5
- 5)
l(7/3e
3'^""'
log 3
>
log
find a positive
lim = (.
n^
In this case, 1
for all
iV^^Mj
Choosing
If for
n>nfj
or
is
(6)
0| <
le/n"
-7
number
n> N,
In this case,
such that
there
3(5
'
6.
2
3
3. iQn
3-10'
a number
is
W>^-f
n>N,
for all
+
+
there
i.e.
e,
XT
^^
3
,.
lm
that
<
-^
<
[c/Wl
>
1^ <e when
Now
p>0
where -0 and
45
i.e.
number
N.
M) such
(depending on
that
00
or
n > ^{M +
1)
N.
should be emphasized that the use of the notations and > for limits does not in any
way imply convergence of the given sequences, since > and -=o are not numbers. Instead, these
are notations used to describe that the sequences diverge in specific ways.
It
if
<
\x\
1.
n-+oo
Method
1:
We
Dividing by logic
Method
which
is
<
a;
for n
negative, yields
n >
the result
is
|a;"|
clearly true.
|a;i"
<
>
0,
we must
nlogio
]a;|
<
Given
when
result.
2:
Let
\x\
|a;"|
THEOREMS
8.
|a;|,
since if
|a;"|
on
Prove that
We must
if
1/(1
+ p), where
|a;|
1/(1
LIMITS
of
if
>
+ j,)" <
0.
By
1/(1
+ np) <
for
all
n> N.
SEQUENCES
lim Un exists,
show that
it
lim Un
must be unique.
- h and
lim Un
= h,
then h
- h.
Thus
1),
lim a"
we have
0.
logioe.
SEQUENCES
46
By
any
hypothesis, given
|m
>
<
li\
we can
n>N,
when
-^e
[CHAP.
\u
hi <
n> N
when
-Je
Then
Ih
i.e.
9.
If
[li
h\ is less
h] =
= A and
lim an
We must
n> N. From
lim 6
hypothesis, given
>
> 0, we can
we have
3,
B)\
(1), (2)
and
we can
10.
for
lim an
Thus
li
= h.
= A + B.
b)
such that
(6-B)|
^E
|(a.
|a-A|
6)
(A
+ B)\ <
for
all
16-B|
(i)
^e
foralln>iV,
foralln>iVi
(2)
Ae
2
forallw>Ar2
(3)
- (A+B)l <
find
le
le
f or all
ti
>
A^
result follows.
bounded.
is
we must show
A,
iV is
-|e
(5),
!(an+6)
where
<
\Unh\
{an
N>0
find
|(a-A)
an-A\
a-Al <
6-B| <
Then from
U\ +
\li
B,
\{an+bn)-(A +
By
+ Un h\ =
\li~Un
such that
< P
\a\
Now
all n.
la|
It follows
that
\an\
<P
for
find
all
|a-A + A|
such that
|a|
< e+
we
if
A| <
|a
for
\A\
choose
|a-A|
for
k>
all
\A\
all
i>Ar,
i.e.,
AT
ai, ai,
.,
on,
e+\A\.
11. If
lim bn
Since
=B
-0,
B = B-6+6, we
Now we
have:
K]
\B\
{!)
\B -
such that
|B-fc|
\bn-B\
<
^\B\
>
|61
^\B\
for
all
n>N.
|6|.
for
all
n> N,
since
lim
K-B
by
hypothesis.
Hence from
12. If
lim On
(i),
= A and
i\B\
lim6
B,
>
|6|
l|B|
for
w>
all
AT.
= AB.
n-+oo
since
lim o
= A
\an(bn-B)
and
ATa.
or
\bn\
la6-AB|
But
<
n-^oo
n-^oo
We
\B\
lim b.
all
<
+ B{an-A)\
B, given any
n>Ni
le
+ |e =
\an-A\ <
e
for
all
>
we can
find
for
^^|g[^
n>N,
(1)
all
n> N.
^^
where
AT is the larger of
Ni and
CHAP.
13. If
SEQUENCES
3]
lima
We
(a)
= A and
= Sv^0,
lim6
prove
>
find
11
1
_ IB - 6|
\k~B\ - \B\ \K\ <
By
hypothesis, given
Problem
we can
0,
Prom
part
find
find
iB
^^ Ti^i 77^
\B\-i\B
and Problem
lim
n^oo On
lim
-+
T-
Ony
lim a
n-
(a)
(6)
lim
~ ^^
n^5n' + 2n-6
lim } ^(^
3-5/n
lim
n"t,5
+ ^) -
'^^
+ 2/n-6M
n^
,;^ J
_
5
(c)
lim(V^H^-V^) =
(rf)
iim%^i21 =
lim
lim
= A 4 =
B
_
~
,.
3
5
l
0)
+ l/n + 2/n^
(1
w>A/',
we can
~2n
rn
^-^^
^^>
Urv,
2n' - 4n'
3' + n'-10
,.
+
+
2'10
3.10"
number
by choosing
16
81
= (I)'
3
22
+
+
S-lO-'
'
2m
2/w' - 4M
+ lM*-10/n' =
10""
,.
= ir.
made
'
lim
(.>r.l^)*
_ ,.
- i
^^
"^^^
~2
write, if desired,
.-(StI)'
+ 4/-
n-tco
">
0)
1
(V^r+l-V^)^^^5iV^ = H^
^
Vw + 1 + a/m
"-^VJT+T+Vn
2nl ^
4
B
limits.
(1
^^^<i^
n> N
all
41).
3 +
+ +
+ n' + 2n
lim
n-*ooOn
for
'
we have
12,
14.
as
complete.
is
(a)
\B\\K\
(1)
\K-B\
= -
(i)
if
lim ?^
(b)
forallw>iV
'
""*"
11).
Then
=B^
lim 6
= 4,
such that
Also, since
lim
(a)
we can
0,
47
5.)
bounded above,
{m}
is
(c)
is
monotonic increasing
Un
bounded below,
if
m+i
m,
= g^~^
(d) is
m = 1, 2, 3,
(a)
bounded,
....
Now
is
monotonic increasing,
(e)
has a
limit.
[CHAP.
SEQUENCES
48
+ l) +
2(w+l)
which
11,
"
if
"'
llw
10
-7
3n + 2
+5
S en" - llw - 35, i.e. -10
3
true.
is
(d)
&n^
2m
2nonly
-35,
wliich
true.
-4n <
(c)
"
" and
If
Thus by reversal of steps in the inequalities, we see that {%} is monotonic increasing.
Actually, since 10 > 35, the sequence is strictly increasing.
By writing some terms of the sequence, we may guess that an upper bound is 2 (for example).
To prove this we must show that m s 2. If (2to - 7)/(3ti + 2) ^ 2 then 2n - 7 ^ 6m + 4 or
is
(6)
- 5)(3n + 2)
(2n
or
2n-
- 3n + 2
a (2m - 7)(3m + 5),
3{n
is
an upper bound.
a lower bound,
is
is
i.e.
Since the sequence has an upper and lower bound, it is bounded. Thus for example we can write
^ 2 for all n.
Since every bounded monotonic (increasing or decreasing) sequence has a limit, the given sequence
1m|
(e)
has a
16.
A
(a)
2wlim
X 4 3w + 2
In fact,
limit.
sequence {%}
defined
is
2-7/w
lim
+ 2ln
exists.
(6)
Mi-1.
^JZUn,
in (a).
n-+oo
(a)
The terms
The
term
TCth
induction (Chapter
By Problem
m..
x
.
lim w + i
Since
excluded since m s
1/2+1/4+ +1/2""'
lim 3
\/3m2
3"^ + "*,
monotone increasing.
i.e. m is bounded above.
Hence m,
is
bounded
lim VSm^,
= ,._
hm
we have
V^
and
= 3.
(The other
1~1'2"
3
_
_
_
_
3n^<
01
_^
.,
1,-1,1,-1
\,-\,l,-h
.6, .66,
By
(n-l)/10
1(1-1/10"),
lim
.,
Letting x
a;)"
+-
mx
if
+
,
...
...
(-l)"w,
...,{~\Y-'l{n^^),
Prove that
(-l)-',
.666, ...,
Increasing
Monotonic
Decreasing
Limit
Exists
No
No
Yes
No
Yes
No
No
No
Yes
No
No
Yes
(0)
Yes
Yes
No
Yes
(|)
No
No
No
Monotonic
Bounded
Sequence
18.
3"^_ Ma
+1/2
i).
,.
Another method:
17.
3,
1,
0, is
,
is zero).
required limit.
possibility,
is
3'
Thus a
Let X
Chapter
14,
(6)
V'Sitl
31/2+1/4+
3
M2
1,
1).
m+i
Clearly,
given by
is
ux
e.
is
n{n-V)
^
'
^,
x"
+
,
n{n
-^
- l)(w - 2)
g,
..3
x"
w(w
1/w,
04)"=
=
No
n(w
^
+ '^i+
1)
2!
J^
n(n
1),
1)---(m
-,
1)---(w
n+
n\
if-
-'+if(-0-M'-i)(-l
n!
w/V
- w+
1)
1)
1^
CHAP.
SEQUENCES
3]
49
Since each term beyond the first two terms in the last expression
follows that the sequence m is a monotonic increasing
sequence.
It is also clear that
nj
by Problem
Thus
value of
19.
Chapter
14,
= 2.71828.
Prove that
^3!
..
2^
2"-i
lim(^l
+-y =
largest integer
where
e,
then w
x,
n + iy
.-\
a;-* .
a;
in
Problem
lim
lim
The
gm+
and
( 1
n+X )
.^\^
+iY =
( 1
(i.e.
not neces-
18).
(^1
+ -V^
(^1
+ i Y*'
y^'-^n^x)
lim(l+iY''= lim(l+lYCl +
and
follows that
e.
Since
it
-^
-^
n, it
1.
^^
n\
is
w.
2!
an increasing function of
is
i')
e.
INFERIOR
20.
Find the
2)
{a)
2,
1
(a) l.u.b.,
1,
1,
u.b.
= 2,
e
than 2
(6)
g.l.b.
is less
(c)
(namely
e
>
lim inf or
e
(InH),
and
(d)
>
1,
since infinitely
the sequence)
is
greater
many terms
-1, since infinitely many terms of the sequence are less than -1
+ e for any
-I's in the sequence) while only a finite number of terms
are less than -1 - .
(namely the 2nd term).
all
results are
(6) g.l.b.,
shown
(c)
2,1.9,1.8,1.7,
l.u.b.
...,2-(-l)/10
1,-1,1,-1, ...,(-i)-i,
.,
g.l.b.
lim sup
or lim
...
1(1-1/10"),
(d)
-1
-i
...
(-l)m,
lim inf
or lim
..
Sequence
h-h\'-h
Hm =
.6,
lim sup
all I's in
The
(c)
(namely
for any
21.
(&) g.l.b.,
1, ....
for any
c>0
1,
e>0
(d)
1,
1,
none
+00
SEQUENCES
50
[CHAP.
NESTED INTERVALS
22.
a + i
[a, &],
n=
l,2,3,
n=
l,2,3, ...
6+i
a,
6,
there corresponds
.,
and lim
nf
(a
6)
0.
CO
Then ai ^ a ^ 6 g 6i, and the sequences {a} and {6} are bounded and respectively monotonic
increasing and decreasing sequences and so converge to a and 6.
To
shovsr
a=b
that
6-a =
(6-6)
|6-6|
+
+
all
\b-a\
Now
given any
so that
23.
from
(2),
>
\b
find
|6-6]
<
we can
0,
a\ <
Since
e.
any
]6-a|
>
<
+
+
{an
-a)
{1)
|a
- a|
(2)
AT
<
la-al
e/3,
Page
(see
note that
e/3
{3)
positive
we
(6-a)
16.-al
e/3,
c is
result,
a =
or
a=6.
5).
Suppose the given bounded infinite set is contained in the finite interval [a, b]. Divide this interval
into two equal intervals. Then at least one of these, denoted by [ai, bi] contains infinitely many points.
Dividing [ai, 6i] into two equal intervals we obtain another interval, say [02,62], containing infinitely
many points. Continuing this process we obtain a set of intervals [a, 6], w = 1, 2, 3, ., each interval
contained in the preceding one and such that
.
hi-ai
n-+
(6
- a)/2^
- a =
(6
- a)/2"
00
This set of nested intervals, by Problem 22, corresponds to a real number which represents a
and so proves the theorem.
limit point
Necessity.
43.
I.
>
0,
we can
find
such that
\up
-l\
p> N
\u-U,\
<
and
e/2
q>
for all
p>N
and
jit,
<
e/2
\l-U,\
<
l\
for
all
q>N
N, we have
\(u,-l)
(l-U,)\
\u^-l\
e/2
e/2
un + 1,
...
If
lie in
is
we have
lim m
a.
Suppose there are two distinct limit points, say a and 6, and
suppose 6 > a (see Fig. 3-1). By definition of limit points, we have
\up-a\
lla
Then
since
|6-a|
<
<
a)/3
(6 a)/3
(6
for infinitely
for infinitely
many
many
values of
(1)
values of q
(2)
b-a =
(6
Using (1) and (2) in (3), we see that |Mp thus contradicting the hypothesis that \Up - u,\
one limit point and the theorem is proved.
>
m,1
<
(6
have
'*~3~*'
'*~s~*'
1-
^.
g_j
{3)
for p,q
CHAP.
SEQUENCES
3]
51
INFINITE SERIES
25.
infinite series
ar
ar^
^J ar"-^
n=l
(a)
- r)
converges to a/(l
< 1,
(b)
Sn
rSn
=
=
if
Let
Then
Subtract,
\r\
(l-r)S,
diverges
If
\r\
<
1,
(6)
If
\r\
>
1,
lim S
lim "^|
1.
ar^
ar"''^
ar
ar'
ar"'^
ar''
r")
1-r
a{l
f^
\r\
ar
Sn
(a)
if
by problem
7.
26.
Prove that
Since
a series converges,
if
Sn
Ui
+ u,+
...
+ u^,
its
Sn-i
lim Un
27.
Sn-i)
lim S
series
S-S
lim Sn-i
S (-l)"""^
diverges.
"=^
1:
^~'^^"
it
lim (S
iifi
diverges.
^^ ^^'^^
^'
Method 2:
The sequence
of partial
doesn't exist.
**
gums
is
1,
- 1,
- 1 + 1,
- 1 + 1 - 1,
i.e.
i.e.
1, 0, 1, 0, 1, 0, 1,
MISCELLANEOUS PROBLEMS
28. If
lim^^n
Let Un
prove that
I,
Vn
We
I.
Vl
iim
^i
+ ^^+-
^i^i
n
lim ^'
+ V^+ ..+Vn
Vl
+ -^^ +
'
"
+ Vz+ .+VP
^^
Vp + i
^
+
t,p +
if
lim
t;
...
Now
so that
\
vi
+ V2+
Since Jim Vn
0,
P we
|-t>P
it;i
.i;,+
choose
+ z|
can choose
...+^p|
we can
\Vp+l\
After choosing
+^>J
so that
|i;,|
|v|
e/2
so that ior
\Vl
<
+
for
e/2
e/2
{S)
and
(3), (i)
1^1
I
[i;p+2|
'
+ Vn
|i;|
(1)
w > P. Then
+ e/2 _ ( - P)e/2
...
+Vp\
<
(*)
becomes
+ Va +
n> N> P,
+ V2+
n
Then using
kp+i|
<
K^
75"
for
n> N
SEQUENCES
52
29.
(1
+ n + n^Y'" =
(l
1+% + M^
Then
Prove that
The
lim
Un
l],
= -^.
lim m
for
we can prove
-^^ =
If
2|a|
lim
2'
(1
+ m) =
1.
-Un
gP
+
,
"
Mn
6(n" + w)
^
< -^^^^^^-^^
+ w + w")"" =
lim
Problem
(see
1^
(1
We
39).
n>2\a\, and
can assume a
if
we
7^ 0.
call
<
(^^'^Un.
lim m
follows that
it
7),
or
i^)"'"
Un
Un-i
'"''
2'
0.
n*- CO
71-+ 00
31.
l){n 2)
then
1,
^ <
(using Problem
-^^
is
Un + 2
Un + 1
^
<
n{n
a.
lim j-
that
""^
'ul
-^-gj
Thus
0.
constants
all
Un + 1
Un
Since
WM
n(n 1)
Un-i
the greatest integer
i.e.
Now by
0.
Then
n\
+^w(n-l)(%-2)
3!~^'*"
and
a"
M)"
>
i.
where m
+ n + n^
,
+ m
(1
result follows if
Let
[2|a|
,^
30.
+ n + n^)!/" =
[CHAP. 3
The expression
ai
indicated briefly by
a2
ai
^-q: ^-q:
where
+
,
!, a2,
is
ai
ai,
+ ,
ai
-\
0,2
Its value is
when
this limit
-\
as
exists,
and
The successive
terms of the sequence are called the successive convergents of the continued fraction.
repeat after some point, the continued fraction is
In case the constants ai, a2,
Given the recurring continued fraction
called recurring.
.
(a)
Find the
first
A_LJ.
2+ 2+ 2+
at a possible limit.
(6)
Assuming that
=
=
2
2
1/2
^2/5
The
29/12
fifth
convergent
Similarly,
if =
2..0....
From
places
we
is
S=..4U2...,
it is
5/2
,^
the results
2.4142.
=
1
2.5
1
29
12
70
29
'^
~
^
5/2
12
~5
^'^
2.4166...
2.4137...
S = ""^-- ^ = ^-^-.
iS=^-^-
CHAP.
SEQUENCES
3]
P. + 2
any continued
in the case of
+
+
2P+I
2Q.+1
Q.+2
53
(w
+ l)st
convergents
P
Q
Problem
fraction, see
76(a).
Assume
the limit to be given by x. Then clearly we must have x = 2 + 1/x. Thus =' 2a; - 1 =
X = ldz^/2. Since the limit cannot be negative, it must be 1 + \/2
This agrees with the
guess in (a), since ^/2 = 1.4142 approximately.
(6)
or
Note that this continued fraction can be defined by the recursion formula
Un+i
and
lim u
if
a;
1/m,
l/x
Ml
as above.
Supplementary Problems
SEQUENCES
32.
Write the
first
^^
(n\
cos
<{^.
<-'ts^}
Avs
VS_
2x
3=
1^
1!'
33.
^'
8'
5=
'^"'>
4!
LIMITS
of
+ y/5),
a;^
cos
a;
1^
'
a;^
is
|(1
^'
2'^
'
a;^
3^^
cos
'
x^
4a;
+ 4^
(c)
|,0,|,0,f,
...
- VS).
(6)
Show
Ans.
(a)
1, 1, 2, 3, 5,
tS =
cos 3*
2a;
SEQUENCES
Using the
('^^
cos
^"^
7=
(6)1,0,1,0,1,...
1(1
8a;=
'
wa;
''
__L
iT'TI'Tf'
35.
4x'
'
34.
-^
.
2!'3!'
^'^{.r^^i^''^:'!.
<"{<?!;.}.
(*)
1 2-V.=
1,
(c)
lim
(d)
lim
5i^
36.
= .001,
(6) 6
Using the
38.
Prove that
39.
Prove that
40
If
Ans.
.0001.
K| =
lim
if
502,
(a)
then lim m
prove that
I,
lim
(a)
lim
cm,,
Prove that
lim 3"
(a)
If
>
1,
(6)
5002,
<
for
w > AT
all
if
(a)
(c)
lim ul
.01,
cannot be ^.
V^,
= A
lim r"
and
lim {%)""
(6)
<,
lim (f )
(c)
1,
=B #
lim 6
n-+
n+oo
1,
lim ul
(6)
l^,
=V
0.
lim a,
if
any constant,
c is
0.
00
0.
n-*oo
-foo
prove that
50,002
(c)
- l)/(3w + 4)
where
cl
= A/B
a/fe
n-*oo
43.
+ 2)/(n - 1) -
0.
n-^- 00
42.
(3n
lim (2w
lim Vwii
(d)
a positive integer,
is
lim ( l)"n
lim M
where p
(c) 6
such that
37.
41.
[CHAP.
SEQUENCES
54
w-^oo
44-
If
\r\
>
1,
exist.
ni->-eo
45.
Wlim -^-^rr'"'
(6)
lim
Ans.
(o)lim
^ (3-\A^)(V^ +
-3/2,
(a)
2)
-1/2,
(6)
(d) lin,
(c)
(d)
V3/2,
limits.
^^":-^" + ^
.^^Zll'}^!-.
-15,
(e)
1/2,
(/)
ie)
lim (V^iFT^
(/)
(2"
n)
+ 3")"
+ 1)
(a) is
monotonic decreasing,
(6) is
bounded
below,
n2 + w3 + n
47
*'
Tf
"
^^
48.
If
M+i
= Vwn +
49.
If
M+i
= Uun + p/iin)
used to determine
50.
Mi
1,
l,
prove that
p>0
where
+ n'
1
lim Un
and mi
>
=
0,
^(1
exists
and
lim u
prove that
Show how
Vp-
Find the
(a)
(6)
1,
|,-J,f,-|,
Ails.
52.
l.u.b., g.l.b.,
-1, J, -i.
(a)
this can. be
\f2.
mi
+ Ma +
LEAST UPPER BOUND, GREATEST LOWER BOUND, LIMIT SUPERIOR, LIMIT INFERIOR
51.
1.
+ \/5).
If M is
is
and
between
lies
. .
lim sup
(-l)V(2n -
...,(-l)''+'(n
^,-1,0,0
(Iim),
(6)
1),
+ l)/(w + 2),
1,-1,1,-1
(c)
is
...
id)
none, none,
convergent
if
1,
-3,
5,
-7,
...
1,4,1,16,1,36,
+,-"
and only
(d)
if
(-1)-' (2n
...,m' +
none,l,+,l
lim m
- 1),
'-"",...
lim m.
+ ttn,
CHAP.
SEQUENCES
3]
55
INFINITE SERIES
00
53.
2 (I)".
n=l
54.
2 (-1)"-V6".
Evaluate
Ans. 2
''
Ans. 1
o
n=l
ji^ + -J + -J + _1_ +
_.,.,
55.
Prove that
56.
57.
diverges.
\S2n
- S| >
1, giving
MISCELLANEOUS PROBLEMS
58. If 0, g M g 6 for all n>N,
59.
If
jim a
result true
60.
Jim
when
= id +
Let u.
61.
Prove that
62.
If
63.
If
(a)
lim \un+i/un\
<
\a\
1,
lim b
independent of
e is
I,
n,
(-!)},
Jim
>/-
<
1,
\a\
1. 2, 3,
1,
(6)
If
Jim (a +
prove that
S,
m.
n)^/"
lim m
+ ,+
...+
.,
>
66.
lim m+,/m
is
If
69.
If
70.
Prove that
a.
|m|
6 for all
|i;|
71.
Prove that
fining the
72.
73.
Show
lim
> AT and
[a,6],
lim
SJn =
(1
+ l/n)"-, r. =
S 1.]
1,2,3, ...
J-(l
+ Vs).
lim a^
A, lim 6,
B,
0,
where a
...
prove that
A^B.
0.
+ 1^ =
w/
(1
1/re)"
and
(l
+ l/re)-,
is
e.
limit.
()3
+ 2T3T + 2i--
(^)"
i^a^--- -
X.
m/m-i
i(l + 1 + ^ +
\
2
3
number
Is the
1.
that
and lim v
0.
0.
sin 1/n
re
0.
68.
6. sinne)
0.
65.
[Hint:
I.
IS e.
prove that
64.
limit
!.
depends on w?
m-*oo
67.
and lim a
and
0,
= 1+1+1+... +
Let S
[Hint:
+ Vr5)
4(3
i(
+ V^M^)
,e)a
(c^)-l
^...
+
1
...
f + VfTf
1
74.
[CHAP.
SEQUENCES
56
Express
Ans.
174/251,
(a)
Vs,
(b)
(a)
1+ 2+ 3+ 1+ 5+ 1+
(&)
[Hint:
In
+ 1+ 2+ 1+
(6)
(c)
and
\/6,
^"^
2+-
+ (V3-i)
(V3
+ i)/2
less
^^ 25+ 7+
^g_^
:^_^
1) to
obtain
+ iT^vTiy = '-^u^T^
(^f3 + l)/2 (namely 1) to obtain
=
+ (V3-i)/2
V5 +
rf
in
^ 2+ 4+ 2+ 4+
V3
as continued fractions.
3.14159
(d)
+ (V3-l)
\/3
^;^
(namely
^7^
2)
'"-vtl
to obtain
'"-^^^2
75.
'^i
^ ^ ^ '
o^"
>
''*'<'^^
0'
'^^^^
convergent
is
PJQn, prove
(d)
(e)
76.
(a)
OnPr^-l
Q.
(-1)""'
UnQn-l
Qn-2
PnQu-1
fraction.
The successive convergents are alternately less than and greater than the continued
the conThe convergents of odd order are less than the continued fraction but are increasing;
decreasing.
but
are
fraction
continued
vergents of even order are greater than the
The continued
Prove that
PJQn and
Problem
75,
if
then
P+i
Qn+l
P.
Qn
Pn-%,
P-iQ
an+iQl
Ans.
(6)
g
77?
~ Q
(6)
Find the
first
is
ac-
26/15
78.
m+2
Solve the following difference equations subject to the given conditions: (a)
= 1 (compare Prob. 34); (h) m+2 = 2m+i + 3m, Mi = 3, Ha = 5; (c) m+2 = 4m+i
(c) it. = n 2
(6) m = 2(3)""' + (-l)""'
Ans. (a) Same as in Prob. 34,
M2
79.
(a)
[Hint:
(6)
80.
By
Use Prob.
73(a)
(1
yfhr
jTjl
jq:
^^
(i-\/6)'+'
(1
- VB)"
34.]
Work Problems
(i+\/5)"+^
(d)
by
first
(a),
= m+i+^, mi = 1,
- 4m, mi = 2, m^ - 8.
chapter 4
Derivatives
DEFINITION
Let
of a
DERIVATIVE
f{x)
The
(a,b).
derivative of f(x) at
.,-,.,,
fined as
a;
= a;o
is
de-
= iim/(^o + ^)-/(^o)
nxo)
if this limit exists.
Urn
f('^)-f('^o)
A
I.e.
function
x^x
is
RIGHT
f{x) at
if this
as
it
limit exists.
Note that
a;
= xo
- a;o
lira
hand derivative of
fix) at
lim
/(^o
a;
In this case h
DIFFERENTIABILITY
is
is
If a function
in
an
a;
= a;o
However
= a;o
defined as
,
v*/
there.
+ fe)-/(^o)
if it
+ fe)-/(^o)
h (= ax)
in this case
flixo)
y-^'
it
approaches zero.
Similarly, the left
Ax
defined as
is
/(^o
f^ixo)
+ Ax)-f(xo)
must be continuous
Problems 3 and 4).
the converse
f{xo
li^
Ax.o
is
if f'{xo) exists.
and
xo
if
and only
if
f^{xo)
it
approaches zero.
fi{xo).
INTERVAL
has a derivative at
the interval.
In particular
all
a^x^b
if f{x) is
then f(x) IS
a < a;o < 6 and
If
it is
differ-
entiable.
SECTIONAL DIFFERENTIABILITY
A
smooth
function
in
is
an interval
aSxSb
26.
[CHAP. 4
DERIVATIVES
58
GRAPHICAL INTERPRETATION
Let the graph of
below.
The
2/
of the
DERIVATIVE
APQB
f{x)
in Fig. 4-1
difference quotient
QR ^
PR
fjxo
+ Ax) -
fjxo)
AX
^ ^^^
(5)
A^
Ax-0
is
shown
'
PR
As Ax-
0, this
secant
Then
{6)
^^
the point P.
the slope of the tangent line to the curve at
Fig. 4-2
Fig. 4-1
An
given by
f(xa)
/'(a^o)
{X
f{x)
at the point
where x=^x,
- xo)
is
(7)
The
PM
spectively.
DIFFERENTIALS
Let Ax
= dx
be an increment given to
Ay
IS called
the increment in y
f{x).
f(x
x.
Then
+ Ax)-f{x)
If f{x) is
()
first de-
Ay
where -^0 as Ax-0.
f'{x)AX
+ cAX
The expression
f'(x)dX
edx
(9)
CHAP.
DERIVATIVES
4]
dy
gg
f'{x)dx
(JO)
called the differential of y or f(x) or the
principal part of Ay. Note that Ay^dv in
genera
However if Ax = dx is small, then dy is a close approximation
of Ay (see Problem 11). The quantity dx, called the differential
of x, and dy need not be small.
Because of the definitions {8) and {10), we often
write
is
^
^-'-'''
-^^-0
Ax
^^_o A
emphasized that da; and dy are noi the limits of Ax
and Aj/ as Ax-^0, since these
limits are zero whereas dx and d?/ are not
necessarily zero.
Instead, given dx we determine dy from {10), i.e. rf^/ is a dependent variable
determined from the independent
variable dx for a given x.
It IS
Geometrically, dy
by the
RULES
If
1-
~{m+g{x))
^f{:c)+^g{x)
2-
^if(^)-9{x)}
^f{x)~^g{x)
3.
d^{Cm) =
4-
a^imai^)}
^mX
^(-)^/(-)-/(-)^^(-_)
,.
li
The
cr{x)
where C
+ g{x)^f{x)
f^^)j^9{x)
if
f{u)
results {12)
where u =
dy _
dx ~
and
then
f{x),
If
f{t)
and y
dy
dx
.du
f^d^
g{x)}
d{f{x)g{x)}
df{x)
iffl'(a;)^0
and v =
dy du dv
du' dv' dx
g{v)
h{x),
(12)
then
(^^)
__
dx/dy
i^^)
then
rfj//df
dx/dt
dg{x)
f{x) dg{x)
/(.),.(.)
= f'iama'ix)
^
~
f'{t)
f'{x)dx
g{x) df{x)
(iS)
^'(i)
g{x)f'{x)
f-\y);
g{t),
f{x)g'{x)
.(.)n.) -
_
-
dy
dx
8.
any constant
is
^^^^
valid.
f'{x)-g'{x)
composite functions.
7.
= xo
f'{x)+g'{x)
[5,(^)]2
f{u)
Similarly
C-^f{x)
<^a;[^(a;)/
6.
DIFFERENTIATION
for
/,
is
line
For example,
g'{x)dx
{f{x)g'{x)
=
+
{f'{x)
g'{x)}dx
g{x)f'{x))dx
[CHAP. 4
DERIVATIVES
60
DERIVATIVES
SPECIAL FUNCTIONS
of
1-
4(^)
2.
^z."
'
du
n"-^
<?
du
dw
du
secittantt^
,
'^-
d^^^'^''
-CSCitCOtM^
8.
j-cscu
9-
d^l^"^
10-
d^^OgeU
11.
d
= aMna^
^a"
12.
^
d
4^
dx
du
loga e dw
^ n =i 1
= -ITd^ a>0,a^l
d
^mU
17
du
1
^
d
J sec ^u
^/^^3ida;
dx
18.
d
dx
19.
d
T smh u
dx
du
^^da;
20.
-y-
cosh u
smh M -J
21.
tanh u
Tdx
22.
f cothtf
dx
23.
-V-
d
sech u
dx
- sech w ^tanh u
24.
d
V,
u
csch
T
da;
- csch u coth w
+ u^dx
dM
-^^
dx
du
da;
'^^^"""dx
csch^u^^
'^
sinh-^M
cosh- lit
27.
e" -T-
J- tanh" ^M
dit
28.
5 coth-l^t
14.
f^
J- COS
dx
d
du
1
'
-jz
dw
y/l-u^
du
1
'^^
du
1-u^dx'
1-z.^da;'
^
H>^
dtt
-,
dM
u^/u^-1'^''
1
30.
. -,
T-sech^'^i
da;
"'^
l^l<^
1
29.
^^
du
da;
da;
..
Vm^ 1
du
dx
e"
f
da;
^l+u'dx
dx
26
du
2E
f-ifM>l
\ + irtt<-l
dM
1
fi^
+ ifM>l
l-itu<
f
^Vw2-1^^
.
<^^
d
^ 1
j-cot"*
dx
.
r,
du
16.
1.
2.
:j-csch-iii
da;
M^M^ +
dtt
dx
an
If fix) is differentiable in
where y = f(x).
,j"
f.(r\
\Jt'),
y
or
oi
by
2/'"^
the
/'"H^)'
first,
i-?^Vf^.
dx^
dx\dxj
or
0,
second, third,
Similarly
where
.
is called
Thus derivatives
....
of
dif-
^^-^P-
DERIVATIVES
^]
gj
5!!*!'1?^''*"'-
/W-/{o)-0,
^" /^''^^
b-a
Rolle's
theorem
The result
X and xo are in
is
(16)
f{xo)
(16)
with
f'ii){x-Xo)
of the
mean
a+
differentiable in (a,
b),
in
=x
between
which case
If f{x)
and
g{x).
where Rn,
called the
and
a+
Lagrange's form:
n^l^H^
R^
Cauchy's form:
i?
...
+ f">(aKb-a)" ^ ^^
/'"-"(I) (ft-a)"^^
a^f<;o
+ !)!
/'"^"() (&-^)"(6
+ nx,)(x-xo) +
/^"^(^ o)(a;-a^o)"
w!
This
-a)
(i
^^^^
(^U)
^^
x and xo
between xo
if
/<"^"(g)(x-a;o)"^
hm Rn =
Note
(%
IS
If
(18)
Xf
f(xo)
where
and differentiable
(i7)
are continuous in
such that
'^^^
^^^^ '^^^ '^^ written in various alternative forms; for example,
^f "xl*
are in (a,b),
then using Lagrange's form of the remainder
R, we have for
fix)
^/i,
/'(a;), flr'(a;)
SLXlCx
a;
a<i<b
WW)
a;o
yields (16).
+ r(a)(b-a) +
/(a)
is
+ !)!
V'<''<'J
used to approximate
about x
= x,.
If
a;o
= 0,
if
(a, &)
'
mean.
is
9ib)-g(a)
4.
h,
and
(16)
The theorem
[a, b\
in
(a, b),
< ^ < 1.
3.
a<i<b
f(^^
if
/(a:)
We
and
and differentiable
[a. b]
(a, h)
[CHAP. 4
DERIVATIVES
g2
SPECIAL EXPANSIONS
The following are some important expansions.
be obtained by using either {20) or {21).
1.
e^
a;
^-3T
1
~
-
a;
oT
2!
qT
3!
^^
+ ;^
n\ +
5T~7T'^'"^'
2.
sin
3.
cosx
+ x)
tan-' X
^-Y~5~^
In (1
4.
5.
In 1-3, lim Rn
-l^x^l.
for
all x.
4T
2T
Y
In
4,
'S
^+
lim
~i
fi
^"
/Y*2n2
(2n-2)!
67
X*
x^
(2n-l)!
in
The remainder Rn
(-!)"-' a;"
y->
-
...
for
is
+ R^
'
'
2^2^^^!
-1< x g 1.
In
5,
lim Rn
L'HOSPITAL'S RULES
If
lim /(x)
lim^
finite,
= A and
is
lim g{x)
for
= B where A and B
co/oo
form 0/0 or
m-
respectively,
If f{x)
xo in
limM
b)
lim/;M
{23)
x-,x g'{x)
x-xo g{x)
If
lim /(x)
00
and
lim g{x)
>,
is
also vahd.
X-^X
X-*Xq
APPLICATIONS
1
that at x
/'^-''(a^o)
f{x)
has a relative
(b) f{x)
has a relative
maximum
minimum
at
at
= l). Then
x = xo if f'-^''^{xo) <
x = xo if /""'(^^o) >
= Xo,
and
f ^'''(Xo) ^
{2J,)
'^^^^-
4]
DERIVATIVES
gg
mmimum
2.
x^xo
point
We
Rates of change.
:tr:^^, It "
3.
J^TJ-^T
s/df^ IS its
d
= fix) must
= f'{x)
''-
'=
--^
If s is the instantaneous
"'^^
^'"' ^*
or
be parallel to the x axis.
*^"" ^"/^^ ^
J'\^
instantaneous
acceleration at time t.
^
"
= f(x)
'S
displacement of a particle
instantaneous velocity and
^*
Solved Problems
DERIVATIVES
1.
Let
3 + a;
= g-^,
fix)
By
A^oie;
Evaluate
3.
points
at all
^^
<.
a;
from the
(3-,,)A(3 + ^) _
nx)
/'(2)
(3
+ ^)
we
Let
/(cc)
/'(5)
(3_^)
Evaluate
/'(5)
= ii^/(5 + fe)-/(5) ^
ft
I/(. +
ft)
-/(.)
6
(3
i,
definition.
ft
= xo, prove
- a;)'
V9 + 2ft-3
^=^
showing that
from the
v^-r^..-a,
+ 2ft-3 V9
2ft + 3
V + 2fc
i;^ V9
"
V^T2ft + 3
..
Then
+ ^)(-l^ _
~
(3-r
"-
3.
(3-a.)(l)-(3
= ^2^;-!.
find
rf^
2.
definition.
/(^o
ji^^^
lim
""
+ 2ft -9
ft(V9T2ft + 3)
9
o
= lim
~ ^ V9T2ft + 3
tL
+ ft)
-/(.,)
^^^^^^
^,^^^^^^
Thus
+ ft)-/(aj) =
continuous at x-x^.
or
lim
/(a;
+ ft) =
/()
_ 11
" ^
= xo.
'"
[CHAP. 4
DERIVATIVES
_.
64
m=
Let
4.
^^0
|q^
Is fix) continuous at
(a)
(a)
By Problem
(6)
/'(O)
22(6) of
= 0?
Chapter
m + h)-m
j^-^^
Does
(&)
i:..
lijw
f{x)
have a derivative at
fW-m)
a;
= 0?
x= 0.
continuous at
2, f(x) is
_
-
limi^^
fesinl/fc-0
^ limsinl
'-
'^
'^
exist.
not have
function is continuous at a point it need
This example shows that even though a
is not necessarily true.
Problem
in
3
theorem
the
of
converse
a derivative at 4e point, i.e. the
but
which is continuous at every point of an mterval
a function
It is possible to construct
s,
J.
Let
5.
(a)
3.
|q^
Is fix) diflferentiable at
by Problem
(6)
x^O
faj^sinl/a;,
fix)
13,
Chapter
exist)
= 0?
Then
2.
f(x)
Is fix) continuous at
(&)
has a derivative
^^(cosi)(-i) +
=
lim
a;
From elementary
Since
= lim(-cosi +
=
conthiuous at
/'(,)
fix) clnnot be
if
sin
= 0?
at
(is diflferentiable)
x=
and
its
value
is 0.
x=0,
i)
-cosi + 2.sinl
(sini)(2.)
2a:
a;
lim cos
l/a=
does not
a;
6.
Present an
fix)
"e,
has a derivative
/'(x)
fixo
at x
= x
+ h)-
if,
fjxo)
given any
^,^^^J
>
<
0,
we
a;
= Xo.
can find
<
^jjen
1^1
<
>
such that
Let
fix)
1^1.
at
= 0. (6) Calculate
Calculate the right hand derivative of fix)
^
at x = 0'!
derivative
a
have
(c) Does fix)
derivative of /(.) at x = 0.
(a)
the conclusions in
(0.
U^^>
(a)
since
since
\h\
|/i|
=h
= -^
for
for
(a), (b)
and
_
=
Hm ^W-^(^
- h+
h
^>0.
h<Q.
(c)
lim
1.-0+
from a graph.
M^
h
1 Ih =
"-*"+
CHAP.
DERIVATIVES
4]
The derivative at
does not exist if
the right and left hand derivatives are
unequal.
(c)
No.
(d)
The required graph is shown in the adjoining Pig. 4-3. Note that the slopes of the
lines y x and y x are 1 and 1 respectively, representing the right and left
= Q.
hand derivatives at x
derivative at
8.
65
Prove that
a;
f{x)
However, the
x^
Fig. 4-3
S g 1.
differentiable in
is
<
<
a;o
a;
Then
1.
h^O
At
At the end
Then
IS
9.
point
a;
= 0,
a;
= 1,
differentiable in
is
customary to write
/; (0)
(a)
8,
y-f(x,)
(6)
As
f'(xo)
in part (o),
2x.
f{l)
fL (1)
line to
f'(l) {x
-1)
or
= 2a;
/'(x)
It
in this case.
at the point
=
2/
where
(a)
= 1/3,
2/3.
j/-i
or
write
/'(I)
= a;^
2/
so that /'(1/3)
f'{x)(x-xo)
We may
1.
and
/'(O)
|(a;--i),
2(a;
= 1.
line is
i.e.
1),
i.e.
2/
2a;
- 1.
6Aa;
a;'
-f-
6a;
fa;
(&)
DIFFERENTIALS
10. If
(a)
(b)
1/
/(a;)
A^
dy
/(
a;3
- 6a;,
+ Aa;) -
find
/(a;)
principal part of
(a) At/,
dy,
(6)
(a)
and
Ay
(6),
Note that
e-0
- #.
Ay
{(a;
-I-
(3a;^-6)Aa;
and dy
(Sx^
i ., that
.w'i^^*
^.'^f
sized
dy and
dx are not necessarily small.
From
A?/
=
=
=
=Bx^~6
(c)
(c)
- dy =
3a;(Aa;)^
as Aa;-0,
i.e.
(Aa;)'
- 6)da;,
(3a;^
- 6)dx,
eAa;,
^^-0
i.e.
since
by
dy/dx
where
as Aa;-0.
definition Aa;
= da;.
3a;Aa;
Hence Ay
(Aa;)^
~ dy
is
an infinitesimal
In case Aa;
11.
is
small, dy
fix)
= ^. Then ^/W+A^ -
^^
^x'^'^Ax
(where
to).
[CHAP. 4
DERIVATIVES
-,
DO
If
= 27
a;
and
Aa;
= -2, we
have
^/27^-V^
\/^ - 3-2/27
Then
or
It is interesting to
^(21)-'" (-2),
V^ - 3
i.e.
2.926.
good.
so that the approximation is fairly
= 25.05,
SPECIAL FUNCTIONS
of
/(-)^^(-) +
{/(.). W)
|:
dx
- -2/27
^(^)
assuming
^^(^)'
and .
are differentiable.
By
definition,
sf,
,x
Mrr.
/(a;
f(x
+ Ax)
.,
/(x)
^.
{g{x
+ Ax -
^,(,) +
g{x)}
r g(x
f{x)g(x)
g{x) {fjx
+ Aa;)-g(x) ]
+ Ax) -
f(x)}
,.^_(^J/(^:^^I:^iMl
^/(.)
,(.)
Another method:
= /(x), = g(x). Then Au = /(x + Ax) - /(x) and A,; = g(x + Ax) - g{x),
Thus
gix + Ax) = V + Av.
MAt> + ^AM + AMAi;
_
(u + Au)iv + Av)-uv
d
- "
lim
T-M-y =
Ax
iiX
A*-*"
Let u
u + Aw,
dx
13. If
it is
/(x
_
+ Ax) ,
A:t->0
/ Av
where
,,
i.e.
r,
Ax-
as
Am
0,
since v
is
Am
*?i
*L
%=%-%
prove that
4-
assuming that
and g are
differentiable.
Ay
as a consequence
Ax^O. Then
where
respectively,
A2/
/(m
+ Am) -
AM
f(u),
ff(x
+ Ax) -
g{x)
AM #0,
let
us write
f^
^y
"
If
we
AM =
define
Ax
e-
^AU
du
as
Am-
and
+ eAM
()
dy
0,
Am^O
For such
= 0.
limit as
(!)
so that
or Am
= 0,
(2)
Dividing
holds.
(2)
by
Ax#0
cases,
j.x.
and takmg the
we have
_
AoA^ ~
,.
"^
dbo
_
~
Ay
dydM.^dM
dudx
Au\
,:^/'d^AM
l^o\duAx ^ 'axJ
dx
dy
du
du
dx
dy
du
Am
j.
A.^oAx
Am
_j_
^^^
^^o
jj^^
a^o Ax
I
(3)
CHAP.
DERIVATIVES
4]
^*- ^^''^^
d^(^"^)
(a)
^(tan:.)
sec^a;,
(a)
fit.n.)
A/^li2^>
and ^(cos^)
^=08^
--(-.)-
(cos a;)(cos X)
^.
~:{sm~^x)
(b)
(sin
.)(- sin
It
sin-'., then .
sin
^
dx
or
"'^
(COS.)
_
"
cos".
^_
cos
VI -
J/
We
''
^^'"'"'^
?unctron':f'r"''
Consider y
f(u)
lim
=
Since the logarithm
^'^
By
log.w.
19,
Chapter
Thenby Problem
16.
Calculate dy/dx
if
Vl-.'
g sin-. S
log(w
^(.2/^)
where
iS.sr--(^)=
.;i.o..(i +
f)""
= m/Am.
;^(logaW) =
^
^.o..
with .
13,
dx^
xy'
(a)
]M1
'
dx
^ xy +
Zx^
y'
=(^2/) +(5)
^(3.")
= dy/dx.
Solving,
4^^")+i(^l-) =
2/
Let
(b)
h,
e^^
+ ylnx ^
cos
cosh,.,
l()
(sinhM)(2)
_
~
=
+
or
(.)(32/y)
(2/3)(i)_6^
Then
(^)(^,)
e-(.2,'
+ 2/)+^+(ln.)2,' = -2
find
(^)(j)^(,
2. sin 2.
."e^
+
+
.j/e^
sin 2..
. In .
l),
where
dy
dx
^'^
2a;.
(6.
(<=2.),
'
cosh(a;2-3a;
2/
2/'
Solving,
(a)
^ differentiable
+ AM)-logM
..)
17. If
(a)
^''
is chosen so that
rather than cosj, =
^/2,
VT^ilH^
is
3,
^
sin^i/
definition,
+ AM)-f(M) _
/(w
x,
^-{i'.(-fy'")
by Problem
^^'^''^
2/.
cosj/^
sin.
_
"~
a;)
cos".
(b)
67
(a)
a."
(^nh.g) =
(coshu)(2.-3)"
and
sinhg+cosh.(f^y
2 sinh (."
- 3. +
1)
(2^
- 3)"
cosh
(."
- 3. + 1)
is
an
[CHAP.
DERIVATIVES
68
18. If
x'^y
y^
2,
find
{b)
(a) y',
(a)
x^
(6)
d
dx
y"
i-(\x^ +
dx
x-1,
Substituting
= -i
2/
=1
and
and
at (1,1)
32/='
-2xy
(y')
^^ '
(1, 1).
+ 2xy + Zy^y' =
-Ixy
(x^
+ W){2xy' + iy) -
{2xy)(2x
+ 6yy')
-1,
;
1:
fix)
Case
Then
in [a,b].
f'(x)
for all x in
(a, 6).
2:
attains
Since f{x) is continuous there are points at which f{x)
Chapter
2).
Problem
(see
34,
respectively
and
minimum values, denoted by
Since f(x) # 0, at least one of the values M,m is not
and that /(|) =
zero. Suppose, for example,
fix)
in
its
[a, b].
For
, >0,
then
maximum and
+ h) S
i^^-0
/(|).
and
If
h<0,
then
f(i
+ h)
and
Fig. 4-4
(2)
But by hypothesis
to the left
must be equal
which case
A
20. (a)
f'(i)
similar
fix)
as required.
in case
theorem.
(a)
M=
Define
Fix)
Then Fia)
Also,
fix)
if fix)
/(a)
(a;
and
(b)
- a).fib)-fia)
and Fib)
m # 0.
0.
specified in Rolle's
the conditions on continuity and differentiability
satisfies
Then applying
F'io
(6)
no
Rolle's
"
f^
^
'
fib)
0,
a<i<b
Then
is
AB
such that
fH).
^ZT^
fib)-
/'(?)
fia)
ACB
AB
fib)
f'ii)
-fia)
a
a<i<b
CHAP.
21.
DERIVATIVES
4]
\.
--^i^lr
4)/(5 - 2)
/o.
(25
22. If f'{x)
^''^^
"
or
^^'
'^'^^^
Since 2
3.5.
<|<
5,
69
the theorem
a = 2, &
+ l0,
2a;2-7a;
states
that
4f
-7 =
is verified.
prove that
(a, h),
= 5.
f{x)
must be a constant
the interval.
Let
xi
<
any two
X, be
f^^'^-fi^^)
X2
Thus
fix,)
at all
pomts of
23. If f'{x)
fix.)
>
<
X2 be
24. (a)
Prove that
(6)
Showthat
Let fix)
(a)
f'ii)
any two
for X2
^-^
J
= tan-x.
<
(a, b),
By
6).
^^'^'^
>
a,
1/(1
tan"' h
Since fix)
+ e) <
1/(1
+ a').
1
>
f'ii)
a;i
^^
- tan-^a <
1/(1
+ a;^)
and
tan~'a
/'(^)
Since
<
h,
1/(1
tan"' &
if
= 4/3
and a
I
25.
=1
<
1/(1
5.
+ e) >
1/(1
Then
6^).
1
1
+^
- a.
Then
since tan"' 1
f+
or
+ 1^), we
tan~'a
b~a
+ 6^
Xi<i< x,,
ib)
strictly increasing.
tan~'6
is
the mean,
Since
x^<i< x^,
From this it follows that if two functions have the same derivative
the functions can only differ by a constant.
a;2
> fixi)
constant.
(a, 6),
^<^')
fix^)
Xl
Let Xi
Then
By
h).
in
we have
,r/4,
J+
== ^^i^l/M
gib)
Applying
- gia)
Rolle's theorem,
f'(i)-ag'ii)
as required.
G'H)
for
or
<f<
6,
<^)
we have
SMn.
ff(b)-gia)
a<i<b
[CHAP.
DERIVATIVES
rjQ
TAYLOR'S THEOREM
26.
MEAN
of the
We
/(6)
m+
= m -
H(x)
in (1)
H(b)
(2),
To obtain H(a)
0.
= 0, we must
_
^4 "
Assuming
then H{x) also
that H'ii)
From
result
(6
- xYA
fib)
(2)
attained by replacing a by
is
H'(x)
= -f"(x){b-x) +
# 6).
a;
choose
f{a)
f'(a)(b-a)
($)
(b-aY
differentiability conditions of Rolle's theorem,
is a value i between a and b such
0.
(2)
2(6
- x)A
A -
for
into (3)
2(&
- |)A =
.
= -f"{i)(b~0 +
and H'(i)
{1).
>
Prove that
sinx
where
| is
Let
fix)
Let
/"(x)
cos x,
51
= 5W180
(a),
x,
since
(x-a) +
/'(a)
and
radians
sin 45
sin 51
45
28. (a)
Prove that
e^
(6)
Prove that
e is
| is
between
'^^
x,
so
that
/(a)
sin a,
2, i.e.,
= 45W180
'-
radians.
Then
a;
-a =
ff/30
radians.
\/2/2,
(cos|)(7r/30)
(V2/2)(7r/30)''
3!
2!
|-(co sg)(7r/30)^
namely
more terms
g^
JL/'-HlV
<-
31^^30^
<
0002
in Taylor's
+ xV2! + x^Bl +
where
-cos
an irrational number.
fix)
/c)(o)
obtained.
^+T~V30y
desired,
is
/' "(x)
x,
i-^,
V2/^\
V^
more accuracy
is
cos 45
Let
3!
= - cos J.
f"'iO
fia)
between a and
If
-sin
J is
/'(x)
= - sin a,
cos a, /"(a)
where
gl
x.
Then
sin X.
Then
(cosg)(x-a)3
a)(.-a)^
(a) to
(sin
+ (cosa)(x-a)
sina
between a and
Use part
f'{a)
(a)
- X) -
Generalizations to n
(6)
(6
f(x)
satisfies
(since
f"(i)/2\
{a)
fix)
is
From
(b)
f(x)
l.
<?<&
where
27. (a)
+^(b-ar
f'(a)(b-a)
and
=
x.
x'
x"
x'''^'e^
/(O)
/'(O)
=
/<\
^'
CHAP.
DERIVATIVES
4]
^^* ^"
"
Ux<0,
^^
(^Tfiyr"'-
all x,
i.e.
-<
e"
'''
^>0'
|i?|
Thus for
i.e.
<
<
%> g
< (^ri)T''
lim
|i?|
lim
|i?|
a;
^'^'^
l^-l
= 0,
If ^
also.
||
and 11m
Chap.
\R,\
3).
0.
xV2\
xVZ\
...
(^^
all x,
^3^^t^'zi::r^'
Choose
l-^-l
,
1
rj-^
''^* -
"^^^^--^^ '^"-^^^^
'-
(i)
by n!
'^^
'-^^*
Then
Now
L'HOSPITAL'S RULE
29.
Weshallsupposethat/(a;)andsr(iK)aredifrerentiableina<a;<6 and
By
m.
f(^)-f{xo)
9{x)-g{xo)
9{x)
^^^^
'*'o^0(x)
__
g'ii)
x"+flr'a)
0/0,
^0,
g(xo)
=o/oo.
(6)
= 0,
where
25),
^ ^
^o<i<x
f'ii)
"
f{xo)
(a)
.i+ff'w
-^
(6)
Wesupposethat/(a;)andflr(a;)aredifferentiablein
a<a;<6, and
where a<a;o<6.
Assume
x^ is
such that
a<x,<x<x,<b. By
^^^^-^^^^
Hence
f(x)-f(xi)
~
f(x)
=,
9(^)
^W
W'({)
lim
^,
f(x,)/f(x)
l-g(xi)/g{x)
x,~
Um
a;
g(x)
i_;+*^^^
->
x.
/^
g'd)
- ff(xi)/g(x)
f (t)
ff'ii)
l-f{xi)/f{x)
= L
oo,
->
'^"
/M
/W =
^'(f)
g(x)
ff(!)-ff{xi)
lim
*-o+
if
and write
"-Al-ZM/ZW; +
(i)
(^)
as
l-Kx^)lf{x) )
(^)
[CHAP.
DERIVATIVES
-72
We
Xi so close to Xo
can choose
that
\f'{i)/s'(i)
<
L\
Keeping
e.
xi fixed,
f(xi)/f{x)
-*+
'"-'''o
(2),
= L =
44
lim
we
fM
lim
Evaluate
e2x_i
^-,
lim
(a)
see that
lim
30.
we
(&)
l + cosTra;
^2
2a; + 1'
,,
->
a;o-,
-> ko,
In cos 3a;
,.
("^
if
In cos 2a;-
.!i?l
lim
,,,
,.
l +
hm ,
(6)
cosira;
lim
.,
ttx
_
5-sin5
-
,.
TT
lim
,.
nm
g'
_
-
cosra;
s
;^
^"^
In cos
3a;
In cos
2a;
_
~
,._.
-3
sin 3a;)/(cos
(- 2 sin
3a;)
2a;)/(cos 2a;)
3 sin
3a;
cos
2a;
^o+ 2 sin
2a;
cos
3a;
jj^^
3 COS 2a; \ _
sin3a;\ / ,.
,.
.llT+^hT^/Uli'JV 2l^^3^; -
3cos3a; \
/ ,.
(^.i, 2 COS 2a;;
/3
\i
^ !
(I)
Note that in the fourth step we have taken advantage of a theorem on limits to simplify
further calculations.
31.
Evaluate
(a)
In tan 2a;
(^^Hm^^^-^
(^)
^T,
hTtiiTS^-
/.
All of these have or can be arranged to have the "indeterminate form"
,
3x
,.
''
- +5
+ 6a;-3 "
(''^
(6)
lim a;'e-
5a;^
a;
lim
_
"
lntan2x
(''^
tan 3x
.11?+ In
6a;
,.
T^
o.
Evaluate
lim
2a;
-^ =
6_
,.
.H?+
,.
2 sec'
2a;
10
lim
= \}}^^
32.
_
"
,.
10^^+6
Vt
_
"
=
2 sec' 2x tan 3x
_
"
!i?JV
3 sec' 3x tan 2x
tan3x\ _
3 sec'SxjV-'?-^ t^^^2^J
^a;
a; tan
^21^(1 + ^)
/2Y
3 sec'
3a;
" VVV^2sec'2^
sin
1-
\xxa
Although L'Hospital's rule is applicable here, its successive use becomes involved. Using Taylor's
Page 62)
theorem of the mean, however, the limit is obtained quickly and easily. We use the results (see
3in^
^_|! +
^,
tan-a;
^+
Qx=.
In (1
+ x) =
f+
fix^
CHAP.
DERIVATIVES
4]
limit equals
(X
li^
33.
Evaluate
lim
73
l + (P-Q)x^
x^lnx.
a:-0
_
~
iol-a;/2d-i?a;^
xHnx =
lim
Hm
i2_
_1/*
,i
lim
The given
to give the
34.
Find
a;
jjj^lncos^
Then
In F(a;)
-sin a;
^^^^
to
lim
^_o2a;cosa;
2a!
- -i. But
lnF(a!)
oc,
;i^^(-sina;)/(cosa;)
'-
=*
Thus lim
F{x)^(cosxY"\
Let
have
'-*"
is
x-^o
'''"^^
- 2a!
sin
= -i
F{x)
(e^-
find
bxf'-,
or
Let
G(x)
\nF(x)
= JLIL^^zM
in (a)
and
and
^hen
(b)
Evaluate
lim'
i-foVsm^a;
34,
lim
a00
- -1
2 cosa;
5xy" =
a;)i/^'
is
1:
applicable.
The required
<
applies.
hm
^7^^^
e""^
We
have
e^
By writing
'^'"^'"''^
a!^sin^a:
it is
seen that
'
Two methods
of procedure
limr ^'-r'^Y-^")
_
_
.-0 \
a!"
^'-
lim F(x).
are possible.
Method
InClimFfa!)!
-*o
rule
x^
L Hospital's
{e^'
lim (cos
(6)
lim G(x)
Then, as in Problem
limF(a;)
lim F{x)
(a)
since
'-*
ln(limF(a;))
36.
a;
Then
35. If
altered so as
We
is
a!
X*
\^
(i^oiiKi;)
"
^-
/Vsin^a:/
Now
lim
'^'-^'"''^
a;*
rule,
[CHAP.
DERIVATIVES
74
x"
s.va?
Iim
= ,.hm
=
Method
2x
2 sin
TT
x_ko
cos a
a;
lo
2 cos
2a;
4 sin
,.
lim
sm 2a;
r~?
4a;'
2a;
fT-j
,.
lim
24a;
x_fO
12a;^
2a;
lim
x^o
4a;''
x-to
8 cos
SI
24
2a;
2:
x^
sin"
sin"
a;"
,.
_
~
a;"
a;
(a;
"(x
x^/3
,.
x^o
a;"
_
~
1/3
xUo
- gV6 + Px'-y
xV6 + Pa;^
+
+
1.
MISCELLANEOUS PROBLEMS
37. If
a;
(a)
= f{t) and
= sr(i)
(a) di//d,
d^ -
d^Tdt
t,
y
1
38.
Let
(a)
iffl'W^"
g'(t)
d/r(
If
= 1/m,
ifi
lim
Prove that
^^Q.
f(h)-m)
^^
^>
lim
r'(0)
Similarly, replacing
/: (0) = 0, and so
lim
h^0+
/'(O)
f'(h)-f'(Q)
- ^ '
Similarly,
lim u/e'
by
,'
/i' (0)
In general, /<*"(0)
Let
fix)
/'"(ico)
(6) /"(O)
0.
e-""'
r
hm
A,
^ 0-
byM-^-, we
and m-^o
find
/I (0)
_
- hm
2u*
/"^'(a;o)
e-i"''-27i-'
,.
hm
n=
for
<
0.
Thus
0.
and so /"(O)
lim l/2Me"
ti-foO
p
2e-"'''
,.
lim
,.
or
0.
1,2,3,
>
if J is
_
-
rule.
..
(see
Problem
89).
at xo according as
By
f'(t)g"(t)
0,
,.
39.
(a) /'(O)
W-+00
[^'(*)]'
(6)
g'it)
a'(t)f"(t)
e-i"''-0
,.
'
U-^OO
/;(0)
_
"
dtVg'(t).
dxidt
WW
1^"'^''
dtW(t)J _
g'fflrW -f(t)g"W l
?wl
m=
f'{0)
dff'(t)\
dx\g'(t)J
d^y/dx^
we have
d ff'(t)\
(6)
respectively.
between xo and
x,
then
= /"(a;o) =
minimum
CHAP.
DERIVATIVES
4]
fix)
f{xo)
- xo) +
f'(xo) {X
/"(")
75
(^-^o)'
+ f"'M{x-x,Y ^ r^'HOix-xoY
3!
2!
=
If
minimum
we have
40.
^""(^)(f--)^
4!
>
/"^'(a;o)
relative
fix.)
4;
f{x)
corridor,
the ladder
is
AB
As
L
i
is
a sec
i.e.,
a sin'
a sec
Then
tan 9
b cos' e
Va^'^
so that
a sec
Although
is
esc e
for
=
fbh,
J 1/3
b cse
is
it
we can prove
length,
tan
b"
6 esc e cot 9
or
,1/3
d'^LIde'^
AB
6 esc e
minimum when
dL/de
minimum
4-
= tan-'v/^
gives
the
by showing that
Problem 88).
this analytically
is positive (see
Fig. 4-G
Supplementary Problems
DERIVATIVES
41.
Use the
definition to
((.)(3a;-4)/(2a;
Ans.
42
T Bt
(a)
f(rv\
'^^'
17/25,
at
41
a;
Tot
-Liet
= 0,
(c)
/c-v^
/^a;;
45.
46.
(h) 2,
(c)
1,
*^ si" I/*''
a;
7^
0,
a;
f{x)
is
{xe~^"''^,
^''''
a;#0
^
Ans.
^^
(a)
a;
*^*
(") /(*) is
= e', show
Use the
results
that
/'(xo)
= e'o
a;
0,
(6) f{x)
_,
(6)
/(x)
(a)
3,
Page
has a derivative
is
continuous at x
= 0,
(6)
has a deriva-
Yes,
continuous at
= 0.
Determine whether
Yes;
point:
(d)
continuous at
| ^^
= 0.
tive at X
44.
+ 3),
lim (l-cos/i)M
63,
(e"
using
- \)lh -
"e, 8
definitions".
1.
sinx,
/'(x.)
cosxo.
and
(a), (6)
48-
Discuss the
Let
g
50.
51.
= 0.
Does
(c)
(a)
and
continuity
(a) 0;
(6) 0;
(6) differentiability
What happens
in case
= 0.
derivative of f{x) at a;
= 0?
(c)
Yes,
of
/(*)
(6)
(d) Illustrate
have a derivative at x
f(x)
Ans.
- l^f^^' 2~*5^.
Ix^ Z, 2<a;S4
fix)
a;
a;
from a graph.
(c)
positive number.
49
[CHAP.
DERIVATIVES
76
where p
0,
is
any
is
Discuss the
and
continuity
(a)
differentiability of f{x)
(6)
in
4.
f(x) at
a;
and only
exists if
aJo
if
/+ (xo)
fL (x).
Prove that f{x) = x" ^ x" + 5x - 6 is differentiable in a S x ^ ft, where o and 6 are any constants.
Find equations for the tangent lines to the curve j/ = x^ - x^ + 5x 6 at x = and x = 1. Illus(c) Determine the point of intersection of the tangent lines in (6).
trate by means of a graph,
(d) Find /'(x), f"{x), f"'{x), r"'(x), ....
Ans. (6) 2/ = 5x - 6, 3/ = 6x - 7; (c) (1, -1); (d) Zx" - 2x + 5, 6x - 2, 6, 0, 0, 0, ...
(a)
(6)
52.
53.
If
/(x)
x^'lxl,
(a) /; (xo)
and
/'(xo+),
(6) /I
(jko)
and
/'(xo-).
= 0.
DIFFERENTIALS
54.
If
55.
If fix)
2/
Ans.
56.
57.
/(x)
+ Sx,
x'^
(a) .0501,
Using
\/36.
If
3/
(c) 5.01,
(6) .05,
Ans.
(a) A^/,
%,
(6)
differentials,
(c)
And
1/x,
Ay/Ax,
(c)
(d) 5,
(c) Aj/
- dy,
{d) (Aj/
dy/dx and
id)
- dy)/Ax,
(e)
(Ay
- dy)/Ax,
(6) 0.12,
sin X, evaluate
(a)
Ay,
(6) dj/.
if
a;
=1
and Ax
.01.
(e) .01
dy/dx.
(e)
(c)
2.0125
(c)
- d2/)/Ax
->
as
sin 31,
(a)
Ax
(6)
In (1.12),
0.
Prove:
^"^
59.
Evaluate
'
dx
tanhii
Compute
f^n^) + fx^ix),
(6)
^{/(x)
p(x)}
J^fi^^)
'
^s(x),
'^"'^'^"'
[g(x)r
j^ {x"
3 In 4, (6)
g(x)}
In (x^
- 2x + 5)}
at x
= 1,
(6)
at x
= 0.
|V^
(c)
61.
(a)
(a)
dx\gix)j
Ans.
60.
f^{fix)
ia)
(a)
a"
sech^M-^ where m
dx
(a)-^tan-'x,
dx
a differentiable function of
(b)^csc-'x,
dx
(c)
-^
dx
sinh"' x,
(6)
^cscm =
^;
x.
(d)^coth-'x,
dx
-esc m. cot m
CHAP.
62.
63.
''
65.
Uy-x'.comvniedy/dx.
If
2/
67.
Calculate
If
If
2/
69.
dy/dx
(a)
and
tan
sec
a;
Aws.
68
+ 2)}''-'<^'+=>,
{ln(3..
Prove
(a)
2/
(6)
-1,
dx"
Establish formulas
y'"
tan
(c)
_
"
t^'i/
^^
tJiQf
and
V2,
= '^^^' ^^^
differentiating].
Ans.
^^^^
Ans. x^X
V
(j^ + ^
a-g- -Ling
'" '" ^
\nx).
2)^/
/.
. and . are
a;2/^l
+ 2/2)(l + ^y^),
2(1
evaluate
t,
:= 0.
d'y/dx^ if xy -\ny = 1.
- 2a;2/V(l - a;?/)^ provided
(6)
prove that
a;,
dy/dx at *
find
^"' ^
(Sj/'
(j)
77
[Hint:
= ^^^^
LLlSle!''^^^ "
Atis.
66.
DERIVATIVES
4]
(a)
d2//dx,
d'y/dx\
(6)
(c)
d2//da;',
at
= :r/4
3V2
d^a;
(a) 7,
dx\
/(
\d^J
~dyy
and
18
(6)
'
on Page
27,
(c)
it
holds.
60.
71.
^^'
~^- 1"^
aLn.V
fapplicable ?
to TJ-^'
this ;
function,
"^
i'nof
IS
not
m?
i.e.
x'{l
no value
is
- x)\
S g
a;
"''*''"'
^*r,?,*''^
[Hint: ^A^*
Apply
Rolle's theorem to the function
73.
0<a<6,
(a)
If
(b)
Use the
prove that
result of (a) to
74.
Prove that
75.
76.
77.
78.
(:./6
If f(x)
(6)
(a)
(6)
Prove that
at all points of
(sin
(a)
By
placing a
Prove that
81.
and
of the
In (1
<
(a, 6),
=
a;
= ^ *h^^
- sin
In 1.2
(./6
'
-1.
< (b/a-l)
< 15'
by using the theorem of the mean.
1/8)
is
20(6).
monotonic decreasing in
(a, 6).
cot|,
in (a),
where
show that
W2).
is
(6)
Prove that
between a and
= a;/2.
sin x
2a;/;r
f or
a;
S ^/2.
6.
if
a;
<
0?
MEAN
a.)
'
cc
^2
^3 ~ ^
+
4
Ans.
(6)
(c)
cosh
(d)
0.741
1,
="'
5(1
+ j)5
'
0<,<.
" < ? < *
/m
(o) tt
Use
(o)^ .
to
.
evaluate
(d)
e~'-
to 3 decimalplaces.
a;.]
In 6/a
*o <
^^^^^I^^^^
In (1.1)
80.
a;)/a; is
.6
e''
1.
paragraph of Problem
is f{x)
Prove that
TAYLOR'S THEOREM
79.
in the last
- a/6) <
< sin-
V3/15)
(a)
(a)
(1
show that
C'^^t'^'=t
there
= 2,
>
(6)/
\
=3
(c) w
vj
anv
any
82.
where
m-
f(x)
is
- ^^^_
f'{x)(b-x)
-,
^)^
(6
= 0.]
where
respectively,
84.
Derive the result (21), Page 61, for the Cauchy form of the remainder.
[Hint: Apply RoUe's theorem to
H{x)
83.
[CHAP.
DERIVATIVES
rjg
By
writing
/i
<<
and
fc"+'(l
-)"/'"'"'(
+ gfc)
1.
obtain Taylor's
in place of the last term (6 - x)A in the hint of Problem 82,
special values
the Lagrange remainder, (6) the Cauchy remainder, using suitable
- xYA
(6
theorem with
(a)
for p.
RULE
L'HOSPITAL'S
85.
s'"
,s
(6)
lim
__../
cos 3a; - 2 cos 2x
'^
lim
(d)
lim x^e-^'
Ams.
(o;^
(a) ^,
[n)
9 4-
_Z + cos
-,
lim
-1,
(6)
e-"\
(o)
e\
(c)
a;
-4/,r,
(3^
(/)
"
-o
- 1) tan ^x/2
(c)
lim x^ In x
(e)
2*
limits.
lim {llx
(i)
2')/x
(j)
Im
/a +
/
(g)
lim
(1
- S/x^
(k)
lim (1/x^
lim
(1
+ 2.)-
(I)
Urn'-^^^lE^Z^LlR
(d) 0,
(e)
0,
<
(n)
x--"
==-^o
(/)
In 3/2,
(g)
e-\
cot^ x)
(h) 1,
(t)
0,
(j)
,.
lim
^^
esc x)
/^sina;\''^
(^-^
x
-o\
+ e'+
(o)
lim (x
(,)
Jhn^(six)-^
(fc)
1,
(0 i,
|,
e^'Y'^
(m)
6,
[p) e
MISCELLANEOUS PROBLEMS
86.
Prove that
87.
If
A/(x)
(6)
89.
90.
91.
f{x
+ ^x)-f{x),
(a)
if
if this
0<x<l.
A{A/(x)}
prove that
/'"* (x)
88.
is
any positive
^.'f{x)
integer,
+ Ax) +
^^
f(x),
A" f(x)
"
(^x)"
40.
f(x)
x",
x>0.
= e~^.
and
PAQ
2 f(x
show that
(c)
for
MO) =
If fix) is the function of Problem 38, prove that
Taylor's series with a remainder for this function and prove that f{x)
cannot approach zero as w^ and discuss the consequences of this.
(a)
path
2Ax)
limit exists.
f{x
where
is
such that
vihi
ti
1,2,3, ...
(6) Write
(c) Explam why K
Rn.
P ""v
N^,
Medium
velocity i
CHAP.
92.
93.
94.
DERIVATIVES
4]
79
A variable a is called an infinitesimal if it has zero as a limit. Given two infinitesimals a and we
^8,
say that a is an mfinitesimal of higher order (or the same order)
if lim a/ji =
(or lim alp = l= 0)
Prove that as a;-*0, {a) sin^2a; and (1 - cos 3a;) are infinitesimals
of the same order, (6) (x^ - sin^a;)
IS an mfinitesimal of higher order than
{a; - In (1 + a;) - 1 + cos x).
Why
lim^^-^^J^ =
Un
m'e-"',
Explain.
95.
[Hint:
Assume
96.
hf'{a)
isa+h,
If
a;
4a;
f(a+h) =
so that
to 3 decimal places.
Ans.
Dy =
(a) 3.268,
(6)
98.
99.
If
f'(x.)
borhood of
100.
-^ (x^ sin x) =
Prove that
{x^
n{n
- 1)}
= ... = r-^xo) =
- x. The point xo in such
f"(x,)
a;
>
'
given by
i in (a, b)
but
case
{a, b)
such that
sin (x
+ w^/2) -
a;'
- 2a;^ - 2a; - 7 =
0,
...
d'-y/dx".
Xr(D''-^u)(D'v)
Chapter
...
uD^v
1).
2wa; cos {x
+ nW2).
/<^"+"(a;)
is
|/"(c)|
(a)
1.131
D-(uv)
where
0.
is
'
approximately.]
97.
f{a)
= 0, show
2) ?
2 3
1
'
f(a+h)
w=
Chap.
^^A_
/'(a)
{/(ft)
f'(b)
_ /(^)j.
0.
(jj^^
chapter 5
Integrals
DEFINITION
of a
DEFINITE INTEGRAL
the area
The concept of a definite integral is often motivated by consideration of
a; = b
x^a
and
at
erected
ordinates
=
the
bounded by the curve y f{x), the x axis and
(see Fig. 5-1).
However, the
i,
xi
definition
fj
Xi
Fig. 5-1
a^x^h
By
writing xo^a, Xn
= b and Xk-Xk-i =
AXk,
|;/(4)(a;fc-a;.-i)
this
/(l)(b-a;n-i)
(1)
can be written
^/djAx.
(2)
sum represents the total area of all rectangles in the above figure.
We now let the number of subdivisions n increase in such a way that each Ax^-^O.
not depend on the mode
as a result the sum (1) or {2) approaches a limit which does
subdivision, we denote this limit by
Geometrically this
If
of
f{x)
dx
(3)
In this symbol
called the definite integral of f{x) between a and b.
integration.
often called the integrand, and [a, b] is called the range of
and b the limits of integration, a being the lower limit of integration
which
call
is
We
upper
limit.
80
f{x) dx is
sometimes
and b the
^^^^-
INTEGRALS
^]
gj^
<^
MEASURE ZERO
A set of points
on the x axis
of intervals enclosing
positive number .).
all
We
is
said to have
can show
measure zero
made
An
condition for
the existence of
PROPERTIES
If f{x)
1.
f{x)dx
fa
DEFINITE INTEGRALS
of
and
is
{f{x)g{x)}dx
Af{x)dx
A C
then
[a, b]
C g{x)dx
f{x)dx
f{x)dx
where
is
any constant
3.
f{x)dx
provided f{x)
is
f{x)dx
dx
f{x)
j^ f{x)dx +
J'f{x)dx
- r
f{x)dx
6.
If in
SXg
6,
m g f(x) S M
m{b-a)
7.
If in a
a;
6, f{x)
J*
f{x)dx
and
are constants,
then
M{b-a)
g{x) then
J
f{x)dx
where
g{x)dx
INTEGRALS
"^^'"^ *^'"'*'"; ^^
iia<b
\f{x)\dx
for
f{x)dx
^^""^ '"
^(^)^^
continuous in
(&-)/(!)
[a, b],
there
is
a point
^ in (a, b)
(4)
[CHAP.
INTEGRALS
g2
r
This reduces to
3.
(^)
f{x)g{x)dx
if sr(x)
(5)
Sf(x)dx
/(^)
= 1.
If g{x) is a positive
in
such that
[a, b)
f{x)g{x)dx
a point
is
sr(b)
(?)
/(a;) da;
4 Generalized second
C"
fix) g{x)
dx
we
if
gia)
If f{x)
fix)dx
gib)
by
replace continuity
()
fix)dx
integrability.
INDEFINITE INTEGRALS
If fix) is given,
Clearly
Fix)
+ c where
c is
Fix)
if
[F(x)
+ c]' =
F'(x)
Thus
/(x).
all
in-
We
by a constant.
is
Example:
If F'(x)
tive of
= x^
F{x)
then
dx
i x'
= xVS +
is
x'.
FUNDAMENTAL THEOREM
of
INTEGRAL CALCULUS
and F(x)
fix)dx
Fib)
fix)
[i.e.
F(x)
(9)
Fia)
the definitio
-.
whenever an
Example:
To
calculate
indefinite integral is
f xUlx,
C x'dx
we note
F(2)
it is
that
known.
F'{x)
F(l)
= x\
F(x)
= xys + c,
f-)-a"-
and we have
= i
<^HAP.
INTEGRALS
5]
83
by writing
^
J
It follows that
^ f{x)dx
f{x)dx
f ^(^)^^
d^J
(^10)
/(^)
{11)
For
etc.
The
^^'^'^'
d^X
f iiElj^j
dx Ji
variable.
f{t)
dt
We
use any
can write
^^^^
(11)
(^^)
x^
of
f{x)
dummy
we can
Thus we have
Example:
CHANGE
variable upper
VARIABLE
of
If a determination of
_ ^n^d(x)
dx
2 sin x'
dx
sin x
INTEGRATION
fix)
dx
not immediately
X f{x)dx
where after obtaining the
terms of
x, i.e.
//{fir(i)}sr'(t)cJi
indefinite integral
{u)
Page 59).
The corresponding theorem for definite
y
where
fix) IS
f{x)dx
integrals is
J^
f{g{t)}g'{t)dt
{15)
INTEGRALS
of
SPECIAL FUNCTIONS
The following
an
identity.
be added.
^"^^
^-
3.
J IT ~
^
smudu
J
^^pi
^^-1
^"1^1
4.
^-
-coszi
6.
J
j
cosMdM
sinM
tanwdw
=
=
=
fcotttdM
In |secM|
In |cosm|
InisinM|
[CHAP.
INTEGRALS
84
X'
("
7.
=
=
sec udu
9.
10.
tanw
csc^Mdw
-cotu
jsecii'udu
taiihu
C cscti^udu
,.
-cothM
7r/4)|
22.
-'
cot^l
sechutanh udu
-sechu
f cschucoth udu
-csch^t
23.
^4.
Jsectant.d
11.
ln|tanM/2|
jsec^udu
In |csc%
21.
tanMl
In tan {u/2
CSC udu
ln|sec%
^^
25.
sin-i- or -cos-^-
/^
^6.
Inlii
^/vFa^
r
du ^
C
^
\
=
^-j
2
I
2_o,2
sec
+ V^?^^
12.
JI
cscucotudu
-cscM
on
27.
"
u
1
1 Xtnn-i
Trtan-i
+ a^
14.
e"
I
15
-'
^
C^tanh M dw
^
17
30.
coshudu
du
^=
29.
13
16
u^
d.u
dw
An
,
sinhM
-In
_^u
a ... \1 ..-i
\
1
_,,0,
77^08-'- or -sec ^a
a
M
a
= %y/u'd
^
^2
,
r-j
ri +
f yoj^-u^du
= u2^Va2_^2
Inlsinhtt]
'
'
32.
\/i?T-^
g-ln |w
Jcothwdtt
a.
Jfv^^^^^dM
"
31.
^u
,
In cosh
i
^
Ma
M+a
1^0^^^^
-^
**
1
^
+-1,u
cot
--cot
or
*^
,-!1
_,sin
19
sechwdM
tan-i(sinh%)
*^
20
33.
hu du
gau cos
5^
e^/^cosbt^
\
of
(^^
bu
sin
e"" (a
b cos bu)
WV^
=^
+ bsinbM)
WV^
INTEGRATION
Integration by parts.
^ udv
where u =
uv-^vdu
f{x)
and v
Partial fractions.
is
= g{x). The
Any
2:
P(x)
less
form
in
^^^.^
Ax + B
:^^2 + {,x + cY
^'^^^^ r
_
-
sum
o Q
1,^,6,
of
A
-3
3a;-2
_
~
(4* - 3)(2x + 5)'
+2
5a:' + 2a; + 4)Ma; - 1)
"^
{2a;
4a;
_
""
a;
(a;''
^^
rational function
Example
j f'{x)g{x)dx
,
1:
f{x)g{x)
,
Example
^ i{x)g'{x)dx
or
the interval
tinuous derivatives in [a,b].
[a, b]
2.
34
SPECIAL METHODS
1.
e"" sin
-coth-i(coshu)
fcsch^id^*
+g
+ 2a; + 4)'
Aa;
(a;'
g
+ 5^
C
I
(2x
+ Z)
+ 2a; + 4
Ca;
a;'
+ 5F
D
2a;
+5
g
x-1
The constants, A, B, C, etc., can be found by clearing of fractions and equating coefficients
methods (see Problem 21).
of like powers of x on both sides of the equation or by using special
CHAP.
INTEGRALS
5]
85
3.
4.
22).
IMPROPER INTEGRALS
If the range of integration [a, b] is not finite
or if f{x) is not defined or not bounded
at one or more points of [a,b], then the integral of
f{x) over this range is called an improper integral. By use of appropriate limiting operations,
we may define the integrals
such cases.
Example
Example
f^
1:
f^=
2:
Examples:
f'
Jo ^
Um
f j^
lim
f'^
li,
->0+
f'
J^ X
limtan-x
=
=
lim
Hm+
2Vi
In.
=
=
-..0
we
lim tan~'Af
7r/2
lim (2-2V^)
lim ( In
e).
(i.e.
converge).
Chapter
does not
in-
12.
NUMERICAL METHODS
for
yf{x)dx
Aa;{2/o
The geometric
2.
2/i
+ 2/2+...+i/-i}
interpretation
is
or
Ax
{y,
+ y, + y,+
..
+y,)
^u)
Trapezoidal rule.
f{x)dx
(i7)
Simpson's rule.
yf{x)dx
This
(I.e.
is
^^''
{18)
corresponding to
the curve y
4.
= f{x) by
used, as in
Problem
26.
[CHAP. 6
INTEGRALS
gg
APPLICATIONS
The use
Solved Problems
DEFINITION
1.
DEFINITE INTEGRAL
of a
prove that
2.
Express
Let a
limi
= 0,
^
^ Jk/(
in
lim^i/(a + ^^)
limi/(|.)A..
(a)
Problem
f x^dx
Express
definite integral.
(a)
If
fix)
= x\
(6)
1.
- i
/f-")
then f(kln)
(fe/)''
fcVw^
Thus by Problem
i^
n(n +l)(2w
+ l)
6^?
_
~
2,
^da;
1,
(1
,.
+ l/w)(2 +
1/n)
/*
virhieh is
Note:
By using
1V3
0V3
limj^
The required
we
observe that
1/3.
Evaluate
dx
/(*)
4.
given
as a limit of a sum, and use the result to evaluate the
(6)
^^
Then
lim
(a;V3)|J
/(-)
as a definite integral.
lim
3.
^j^pg
limit can be
= x\
"
"
"
a;
n + n\-
vifritten
Jo
using Problem 2 and the fundamental theorem of the calculus.
a;
=1
is
equal to i.
x^dx
J^
^
=
CHAP.
INTEGRALS
5]
Prove that
5.
Iim;^jsin|
n-foo 71
Let
a-0,
= t,
71
f(x)
+ sin +
71
sin
71
cost
Then
1.
"2 sin
Mw
^'""^
inv^,
+ sin^^?-^l =
...
Problem
in
87
sin
a;
da;
l-cos(
and so
lim
n-oo
^^^ =
lim
i
n
=l
COS
0.
MEASURE ZERO
6.
25(a) of
Chapter
PROPERTIES
'
3),
of
''
set
=^^,<c..x,,x., ...
^'^
iZ
DEFINITE INTEGRALS
''"'^""""^ ^"
f"^'
proJe'thal
^^
m(6-a)
mS
^"^
/(a.)
/(a;)da;
g M where
and
g M(&-a)
(6)
(a)
We
(a)
geometrically.
have
mAx
Summing from k =
to
^^Ax,
/(U^**
MAX,,
l,2,...,n
(x,-a)
(x.-x,)
...
(6-a;-0
= 6-a
follows that
it
<b - o)
Taking the limit as
the required result.
(6)
Assume
shown
f(x)
/(4)
and continuous
in [a, h]
Aa;;c
S M(b - a)
yields
with graph
It is geometrically
evident that
Area
ABCD g
Area under y
i.e.,
w(6-a)
A
fix) IS
/()da;
M(b-a)
made if the
removed. The result also holds
sectionally continuous in [a, b].
restriction
II
f(x)^0
is
Fig. 5-2
are constants,
8.
[CHAP.
INTEGRALS
gg
Prove that
By
f{x)dx\
inequality
Page
2,
\f{x)\dx
Ip'^sinnx,
Then
aj^
P^^^^d^l
liml
If
(a)
/(a;) is
by Problem
I*
Jo
[a,h],
/(a;) da;
[a, h\,
we can
;. ^(^)
^^
and
such that
w g /()
^ M. Then
and
(see
Chapter
2,
Problems
34, 93).
/() da;
by
a.
such that
in (a,b)
(b-a)/(|)
If fix)
a point
is
find constants
m s
(b)
2w
geometrically.
7,
The required
dx
INTEGRALS
for
(a)
continuous in
n^
<
+*^
r
(ft)
p'^l sinwa:
Jo
0,
continuous in
/(a;) is
we have
-^ 0,
w"
'"'
I==l
|Jo
a<b.
if
3,
1=1
7(6),
we can
interpret
j^
f(x)
dx
as
= f(x). Geometrically this area should equal that of a recthe shaded area under the curve y
for some value | between a and 6.
height
6
and
a
base
f(0
with
tangle
FUNDAMENTAL THEOREM
IL
If
F{x)
C f{t)dt
Fix
by the
first
Then
+ h)-
where
Fix)
if
is
of
any point
INTEGRAL CALCULUS
f{x) is
C
-\\
/(^)
f{t)dt
^
[a,b],
prove that
-.
mdtV
"I
F'{x)
f{x).
/-' + "
f(t)dt
dt
fit)
-^J
interior to
F'ix)
continuous in
x+h
^
^
li_>0
[a, 6],
F(^
+ /^)-^W
ft
Hm/(|)
/(.)
''-*''
since / is continuous.
If
=a
cases as well.
or
= b, we
^^^^-
INTEGRALS
5]
gg
12.
where
any constant
c is
= c,
Since F{a)
13. If fix) is
it
follows that
continuous in
is
any point
and F(x)
is
continuous.
li
If
at X
a;
a,
= a and
and x
a;
= 6, we
F{x
dt
/(t)
prove that
[a, b],
lira
Chapter
22,
i?'(x)
we can
or
j' /(t) dt =
4).
F(a)
write
^a
Problem
F(h)
\a,h\,
interior to
r mdt +
F(x)
f(x),
f{t)dt
is
F{b)
F{a).
continuous in
[a,b].
+ h) -
= limhm) =
"-*"
F(x)
is
continuous
o.
Another method:
contifuou^s'"*'^^'"
CHANGE
14.
"
^""''^'"
^'"'^
VARIABLES
of
and
F(x)
dF =
Then
f{^)dx
/(a;)
dx,
^^^^^^"^ ^'
^'
SPECIAL METHODS
83, for
and
dG =
of
INTEGRATION
G(t)
i{g{t)}g\t)dt,
where x
J^
p(t).
f{g(t)} g\t)dt.
'**"^^t^^*
FW
^'
/(^^'^^
="G(t?+I'''^*^'^*'
i^()
Fi.)lZ^'l:^xJb=:.Z^^r..''^^^^
r
f
C
f
/(a;) da;
f(x)dx
dFW =
/{i/()}^'Wdt so that
()
O,
dG(*),
so that
from which
= .
Hence
f{g(t)}g'(t)dt
as required.
15.
Evaluate:
r(a;
(a)
(6)
+ 2)sin(a;2 + 4a;-6)da;
^^
f'
(c)
J*
^2*^)^^
^'^
r2-nanh2i-dx
(^^
X -"
sin
- ^C 7!^^^^
^ a;^
,o\
J-.V(x + 2)(3-x)
Jo
*^
(a)
Method
Let
;^
J
(6)
(a;
+ 4a;-6=M. Then
J
sinwdit
(2a,
+ 4) da; =
dzt.
--cosm +
+ +i
a;
(a;
becomes
(a;'>
2:
+ 2)sin(a;^ + 4a;-6)dx =
Let In
Va;'
1:
Method
^^^
(/)
a;
M.
Then
(da;)/a;
cotMdtt
sin
(a;^
+ 4x - 6) d(a;^ + 4a; - 6) =
-1
Inlsinwl
In
sin (In
a;)
cos (x^
+ 4a; - 6) +
[CHAP.
INTEGRALS
Method
(c)
1:
x-i-=u,
Letting
f'
by Problem
J 1^=
Method
as in
dx
= sin"'^ +
= sm-'[^^)+c.
'-(*^)|-, = -'-(l)-"'"-(-|
sin-'. 2
sin-'
2:
a;-i = M
dx
becomes
this
Then
Let
dx
dx
Method
Now when
1.
.6
-1, m
= -|;
_
_
- 1, u-^.
i
and when
a;
rpi, lo
Ihus
14,
1/2
r'
dx
r'
J-. V25/4 - (a;
da;
J-iV(x
iV(x + 2)(3-a;)
=
Let 2'-^
id)
Let
(.)
sin-'.^
du
Then
sm
' X''
r
j
xdx
V x' + x + l
=
lJ(.2 + , +
^/S^
1,
M = tan-'
ri-/
0;
\/3 sec"
2/,
(dx)/x
_
""
dy.
2,
In
l/^/3
u du
[Ssec'uf"
When
= e,
y'
j/
1")"
irin-'
4^sm
^^
-^
144
V^^* + l
l)_|J_^=
- 1 = Vstan.,
Then the
W6.
i
3
= 1; when
becomes
r^g
Let x
\/3 sec'
dy
r_^+j_ ^, _i
-^^
V^T^^Tl
+ .+
Jo
j;,
5/2 1.
"'"'-^'^
i(sin-'a;T
+ i + \/(^T|r+l
u = tan-'
C""^
sin
2.^
l)-.2,(,2
+1 - ^
when x =
u du
IsT^tan^itf"
Let In X
^(^^^r^^^r^
2^
^^,
^%
/,-! ^2\2
X )
i(sm
+ i-i rf,
2j Vx^ + x + 1
1
= -
and 2-'d.
iu'
_,
dx
VisTI^^
'^-^-a/^
^==2^^^
^ C udu
X
- If
sin-'.
- 2'-^(ln 2) dx = du
Then
..
sin-'. 2
^1
-2K
COS
It
dx
V3sec^,.d..
When
integral becomes
dtt
:^
sin
a;
e^
.
j/
= 2.
^^^^-
18.
5]
INTEGRALS
jx" In x dx
Find
if
(a)
-1,
(6)
(.)
91
n = -1.
ln.,
d.
.d.,
n+
^
a;" +
(6)
19.
a;-Mna;dx
In
x d(ln
V2^Tl =
2a;
2/,
Find
a;
Let
ydy
J"
M
In
(a;
udv
+ 3)
In (x
+ 3),
\{\nxY
= y,dv =
+ l).
da;
c.
S> dy;
then d.
= y3'
dy,
In 3
Then
dw
"
'^^
=
a;
= y^"(- + 3)-|/Srl
|-ln(a;
^^1^
^
Jf (a;-3)(2a;
fa; -3 (2a; +
5)
3)
3V(ln
ln3
(hTir
"
a;ln(a;
+ 3'
- T*='
tt
^^"*= on integrating by parts,
= fln(. + 3)-i/(.-3 +
9 1n(x
+ 3)| +
+ 3)rfa= = | -
4 In 4
+|l3
+Z
]dx
6?a;.
fractions.
6- 3)(2a; + 6)
a;
Let
a;
^ "^
^
- 3 + 2ST5"
1:
6-. =
Since this
is
an
(a.
- 3)(2. + 5)
to obtain
A(2.
identity,
+ 5) + B(.-3)
or
6-. = 54 - 35 + (2A + 5).
5A~BB = &, 2A + B = -1 and 4=3/11.5 = -17/11
Then
jjln|a;-3|
Method
SS'-ydy
and we have
3),
InS'*^
- l|^-3a; +
Method
= y dy
/ du
x dx.
Then
Determine
Then dx
j/^
wy
dv
||-ln|2a!
+ 5| +
2:
W, we
22.
...V(n
(n+V)
a;)
Jw +
+I
^^r^ +
J..
(d.)/.,
da;.
/.ln(. + 3)..
21.
a;
Find J*3v2^+^da;.
Let
20.
In
+ l'"""
du
so that
Evaluate
From
^^
i? example, lettmg
""^
i
da;
Fig. 5-3
^'^
3 COS*
we
sin a;/2
see that
u
cos
Vl + M'
a;/2
=
Vl + M"
a;/2
u.
a;
=3
and
a;
-5/2
in
[CHAP.
INTEGRALS
92
Then
cos
du
Also
Thus the
a;
cos^a;/2
sin
a;
a;
M^
fTjT^-
_ 2dM
- i+^-
2cos^a;/2dM
r <^w
J ^^T+J
IT,
integral becomes
.
dx
or
^sec''x/2da;
sin*a;/2
1 tan
m/^
^ +an-'M/2
-r
c
i-
i
i tan"'
(A tan x/2)
v^
c.
J"""
Let
r''_rf(cos^_j
J
i.e.
= irV^-I
Then
TT-y.
cos'
or
-;.tan-Mcosj/r
\/sina
Letting a
>rV2
7rV4.
_
-
aa;
^^
J^
Vsmic + vcosa;
= 7r/2 - y, we have
J'"'^
J/
P&=dx
Then
,
from which 2/ =
7r/2
Jo
V^ini
^I^l
dx
Vsin^ + Vcos*
Jo
\/sin
r"
^
.
and 7
-v/cos
"
jj-/4.
ir/2
_
^
sin" X
J
+ cos" X
sin" X
1
-'
m,
!L
(see
Problem
first
94).
NUMERICAL METHODS
25.
da;
Evaluate
rule,
r^
where the
for
approximately, using
interval [0,1]
is
divided into
(a)
n = 4 equal
Page 85, we
Let fix) = 1/(1 + a:^). Using the notation on
= 1.0000, y.
=
Then keeping 4 decimal places, we have: y. /(O)
(b)
Simpson's
parts.
find
j/3
(a)
= /(0.75) = 0.6400,
The trapezoidal
2/4
= /(I) = 0.5000.
rule gives
f {1.0000 + 2(0.9412)
0.7828.
CHAP.
(b)
INTEGRALS
5]
{yo
26. (a)
(b)
93
Evaluate
As
in
Problem
jr/4
e^^dx
J^
estimate the
is
4,
we
^^^-
Now
"*"
lA
'
4!"
0<l<
+ ~5r
++ +_+ __
a'
+ fy + fT+
^^
+ 5"^ + 7^^ +
IC'^eldx
"^'^
\E\
'
3T
0<i<x'
to 1,
X('
x^,
find
2!
Integrating from
0.7854.
error.
Chapter
Then replacing x by
0.7854.
approximately
maximum
28,
{j.ooOO
5!
<
-
+ ^
fih
9T4T
+ ^
r'h'.^
15!-^''^^
<=
1-4618
r'a'"
= 'X
-51'^^
E =C' ^e^dx)
+ E
e
= 1IT57 <
0.0021.
APPLICATIONS
(a)
adjoming Fig.
in
the
Then
5-4.
Required area
in the figure
shown
2=
lim
"-oo
f{Q Aa;*
fc
lim
=
I
(6)
(4-iDAa;.
(4 -a;") da;
" =/()
^2
Assuming unit
ilfiQAx^.
Then
= Um i ^JiQ Ax, =
=
a;^(4
-x'')dx
lim
= iM
15
2 g (4 -
jj) Aa;.
= 4-{;
29.
[CHAP.
INTEGRALS
g^
j^ Vi
from x =
x^
(dy/dxY dx
y/l
to x
ij Vl + u'
r Vl +
4a:'
dx
l.
(2xr dx
du
l\/5
+ iln(2 +
V5)
Required volume
lim
^yl^x. = ^ f (i-x^dx =
axis.
512,r/15.
MISCELLANEOUS PROBLEMS
and
30. If fix)
are continuous in
^(a;)
for
have
A^
we
31
find C"
(fif(a;)}^(Za;
J^
real values of
all
f^imYdx
(f f{x)g{x)d^
We
[a, b],
Hence by Problem 13
X.
B'
(\g(x)Ydx,
t,
of Chapter
using
1,
with^
C =
{f(^)rdx,
(1)
f(x)g{x)dx
Then on integrating by
parts,
a.
f{x)g(x)dx
r g{x)F'(x)dx
g{x)F(x)\l
g'{x)F(x)dx
\
*y a
=
Case
1:
first
i is
in (a, b)
2:
is
F{i)
g'(x)F(x)dx
0.
^^g'{x)dx
Page
82)
F(i)[g{h)-g{a)]
so that
The proof
C" f(x)g(x)dx
Case
g'{x)
i.e.
(\'(x)F(x)dx
where
g{b)F{b)
g(b)F(b)
F(i)[g(b)-g(a)]
g{a)F(0
g{b)[F(b)-F(i)]
g(a)
i.e.
g'{x)
1.
r
g
0.
f(x)dx
gib)
f{x)dx
we
have,
*^H-^P-
INTEGRALS
5]
continuous in
is
95
[a,b],
[a,b],
dt
(6) Use
theorem
(a)
We
to obtain the
(a)
(see
Page
Assume that
^^dt
we
it
f(a)
f mat
holds for n
dv,
The
1).
f(a)
= k. Then
/-"(*) = u
We
f
Letting
giving the
F(t)
LafirmTiflre
i:
/(a)
-^jf^.
/'' + ( a)(a;-a)i'+i
f(x)
since
f(a)
(i)
du
/<--(*)d*
first
fc
+ 1. Thus
F{t) G(t) dt
F{i)
/<->((), G(t)
= ^^~^,
have
Accordmg
:^
to the
x^
Page
n S
0.
82),
between a and x
(20),
^,
Page
61],
with
6 replaced
by
x.
,_
(21),
Page
61],
with h replaced by
x.
4^^
/'"^^'Wdf
obtain
C -^
lim
M-
we
1,
it
('-)'="'
G(t) dt
Prove that
^*
+ (fe+TyrX
(Ffl)!
F(e)=/^!:!lMi^, ^(, =
Letting
We
/(f)
w=
so that
33.
find
_
-
(6)
in Taylor's
j^^ij'iui
61).
4x^
(^^
+ Zf -
^^xV
(x^
+ 2 + 2x)(x^ + 2 - 2a:).
= _ Ax + B ^ Cx + D
+4
- 2a; + 2
+ 2a; + 2
{A
C)x^+
(B~2A
1
2C
D)x'+
+
=
+
{2A-2B + 2C + 2D)x + 2B + 2D
+
A + C-0, B-2A + 2C + C = 0, 2A - 2fi + 2C + 2i? =
0, 2B + 2i? = 1
simultaneously, A = ^, B = 1, C = -i, D =
Thus
^.
x*
Then
so that
Solving
x^ +
Jf_^_
4,
1
i
f x+2
J x^ + 2x + 2
8j
=
(a;
,^ln(x^
i)
i''*
a;''
r
J
+ sj
a;2
a;-2
^^
;' - 2a; + "
x2-2a;-'
2
(a;
+ 2a; + 2) +|tan-(a; +
1)^
l)
" gj
i (^.
(^rriy^TT ^"
da;
(^:rijF+3
_ 3^ + 2) +itan-(x-l) + C
'^*'";.
We
dx
r"
ij^ rj_i
The
dt
Si sin t^
-j
lim
Evaluate
called
an improper
(M +
+ j tan-' (M - 1)
1)
Let
if f{x) is
continuous in
,^i^^^"^'
sin.3
[a, b]
then
m(b-a) S
l.u.b.
and
g.l.b.
= im.A^. g
is
lim?^^
Since f(x)
is
continuous
M.Aa;.
we can
i.e.
= S ^ M(6 - a)
The sums
fix) in [a,b].
exists.
/(x) da;
J^
f(Q^x.,
78.
Prove that
where
1 ^^_,
ifr^''''
35.
/M!2M + 2 \
X^m'
34.
[CHAP.
INTEGRALS
96
find
such that
W
called the
Now
s'
this
S'
(^)
we
and T respectively.
and
To prove
first
By Problem
s
which proves
89,
S T S
we have
S'
and
s'
S T S S
()
(f).
are
number of subdivisions is increased, the upper sums
to (1) these sums
accordmg
Since
increasing.
monotonic
are
sums
monotonk decreasing and the lower
shall call s and S respectively.
r^a^^rboundTit follows that they have limiting values which we must show that s - S.
we
exists,
integral
the
that
prove
order
to
By Problem 90, s ^ S. In
any
[a,b], it is uniformly continuous. Then given
From
Since
>
0,
it is
(2)
we can
interval
is continuous in the closed
take each Ax. so small that M^-m^
S-s =
<
(M,.-m4Aa;fc
e/(b
<
- a).
It follows that
^37^2^^^" "
U)
positive
and it follows that each term in parentheses is
it must be zero, i.e. S-s.
number
definite
is
a
5
S
since
and so is less than e by (i).' In particular,
and the proof is complete.
Thus the limits of the upper and lower sums are equal
Now
S-s = (S-S) +
(S-s) + {s-s)
'
C^^P-
INTEGRALS
5]
Qrj
Supplementary Problems
DEFINITION
36.
(a)
DEFINITE INTEGRAL
of a
Express
x^dx
j^
integral,
37.
38.
39.
definition, evaluate
Prove that
Jim J
Prove that
^^
/l"
2-
(a)
^ +
3-
(Sx
42.
43.
Work Problem
Prove that
lim J
Prove that
lim
PROPERTIES
i -^L^
45.
47.
48.
If
/(a;)
Property
Prove that
2,
(6)
cos
51.
and
'^
x'
In 2
a;
for
>
1.
/,
/s.
,^0.
is
for
a value
/(a:)
(a)
J"
/() da:
^(a;),
( f(x)dx
prove that
f
s
f(x)dx.
f^Wdx.
INTEGRALS
{^
and
J in
^ e^in ^= cos
|,
in
sin TTX
OSxSir
m S f(x) S M,
Q^xSl
a;
dx,
(6)
(a)^esina:3
^^
S J^-^'V\'^)
f(x)ff(x) <
m 9(h
g(x)
JNow mNow
in
-M
such that
57
C e-
such that
of
mg(x)
then
cos
a;
da:
gin
INTEGRATION
f'^^^^^dt,
C'
(c)
-^1
^'^s.
127-
^^U-.16^-
J/(a;)da;
s-/2.
Evaluate:
jf
'''*
of
all n.
CHANGE
53.
sm X
+l
and
in [a,h]
X
52.
if
(6)9
e".
prove that
(c, 6),
for
ajVs-
Jo
Prove that
a;
...
(a) 8.
f'^^iM^^ s
x+1
49.
te^
Prove that
Ans.
DEFINITE INTEGRALS
of
Prove
46.
44.
(a)
e''
5 directly,
^_l
(6)
C\x--4x)dx.
(6)
Ans.
l)dx,
n-
e'dx
41.
Use the
(5)
(c)
Using the
as a limit of a sum.
(6),rV32,
(c)W3,
(d)
-~:=,
V4a!
-2
(d)
coth y/H
-^
a;'
c.
csch'V^
V^
(e)iln3.
^
**'
INTEGRALS
98
54.
55.
Show
(a)
J^
Prove that
(6)
56.
that
fVa''
Find
^2-'
-ia"
yfof+2xTh -
Ans.
sin"' w/a
In
c,
a;
>
0.
58.
Evaluate
59.
Show that
(a)
cos
a;
(a)
3a; da;,
tan"'
a;^
(6)
da;
a;
a;
e"^^
M = /(a;) and
= sf(a;)
t;
evaluate
Show
that
[Hint:
wo'-'^^dx
x* sin
J^
Use
Mi;<-
(x
a;
ir(x^
(c)
Prove that
dx
-,
Evaluate
r-^r
i "r
i/q
'^
for
-2/9,
(a)
-le-2-{4a;'
(6)
In 2
>
+ m'V"-" -
- (-1)"
M<'i'da;
I
,.,
l)Ma;^
+
.
^,
-,
+ iy
Prove the
66.
Prove Simpson's
(a)
value,
In 2
.,90
(a)
"^
m~
+
x^
^^^
^^"^
(6)
0.6931.
rectangular rule,
(a) U.3ZZ,
Use
1.
(6)
trapezoidal rule,
i.e.
(16)
and
ir/4.
Page
(17) of
85.
rule.
dx
Ans.
(c)
Discuss,
approximately, using
0,10,
65.
0?
+ -rrr
x+ 1
r^rr^
(x
1)
XTT/2 sin^ X dx
a;
a;
sin^x at
12)7^ + 48
assume
i.e.
(a;
NUMERICAL METHODS
63.
jr*
l)
partial fractions,
What
(6)
Ans.
da;.
mV"-^'
A B C D]
62.
c.
Ans.
da;.
- ^ + ^
_3L^
61.
+ 2a; + 5 +
Va:"
+ 2a; + 5
Va;'
(a) If
''
57.
60.
"^^ +
J ^^y^?^ "
^'^
\u-\Ja^-u'
^^^
f
^
du
+ 2^_^y.
f v^?^dM
(o)
- m'
(3
[CHAP.
[O)
X^ 3-v+x
_
i.iuo
J-,
(6)
r' cosh
I
j
x^ dx.
APPLICATIONS
68.
Find the (a) area and (6) moment of inertia about the y axis of the region in the xy plane bounded
Ans. (a) 2, (6) ir^ 4
by 2/ = sin X, S x^tt and the x axis, assuming unit density.
69.
Find the moment of inertia about the x axis of the region bounded by y
Ans. ^M, where
is proportional to the distance from the x axis.
70.
Show
cosh x
from x
to
In 2
is
density
CHAP.
71.
72.
73.
INTEGRALS
5]
Show
xVa'
ellipse
a(tf
sin
y^/b''
'''''''"'''
W3*X'
a^t'To!
(a) If
lemnwcate
p"
''"""'^"^ ^'^
''''''''
cos
e),
^ 9 ^ 2^)
(0
= 0,
and ^ = 0,
Ans. (b)
| f*'
is
a^
p^d^.
is 8a.
is B-a6.
d^xS^,
sinx,
" '^^'^^' - =
"^
a{l
'^
^'^d the
x axis
is
is
/(0)
' *^'
e),
^*'
75.
a;
99
located
by
this
(6)
'^<bi
f Vo^ +
is
(b)
(do/daY
da.
^"
^*:
(h\
^^
T?f^ +,^
MISCELLANEOUS PROBLEMS
77.
[Hmt: Let
78.
f(x)
Prove that
F'(x)
l^
(a)
mean
f'
for derivatives
Page
in (4),
4,
lim
(6)
first
for integrals.
f'^
lim
(c)
6,
f"'
^^
. K
results.
j^^^,
from the
81.]
and
f^ 4^
X'7=
x^e-'dx
80.
Evaluate
fj^.
1+x^'
(a)
(6)
'^
Jo
Ams.
(a)
-^
(6)
(c)
4!
"'"^'^
24
lim
(6)
= -
'^^
f
^^
Xf"' (sin.)-'^*'.
d^;
Xf" ^TT^f^-
(rf
()
3V3
81.
82.
Evaluate
Prove:
^hm^
^ f^yt^ +
83.
84.
Prove that
85.
^-^^^
(13)
+ 1) dt =
on Page
= 0. But
3.
(6)
A J^^^^
,.
dx
4,
87.
Evaluate
lim
""
2. cos .^
cos
.^
- _
^'^
l).
f
^
the .transformation
- -A usmg 4.^
x-l/y.
J_il + a;= ~
J_^ 1+2^2
= tan- (1) - tan- (-1) ="1/4 -(-./4) = ./2. Thus ./2 = 0.
I
Vl + a;'
JV^^^ + ^f^^ +
1
^ V21n(V2 +
f^' ^r^^^f^^
(.)
'
1/2
J,
^,
83.
VlTl^x
(a)
Hence /
^,,. ,/,^
;^57^
2
-
+ V^^T^ l
h
2,0 /^
Ans. |(2V2
,.
1)
88.
Prove that
[Hint:
division
89.
[CHAP.
INTEGRALS
100
f{x)
jj
j^
^^
SlonaT^
[0, 1].
then irrational points of sub(2), Page 80, let |k, fc = 1,2,3, .. .,w be first rational and
and examine the lower and upper sums of Problem 35.]
In
(3)
of Problem 35.
[Hint:
subdivision.]
Assume
[Hint:
90.
91.
S.
prove that
f{x)dx
exists.
[Hint:
discontinuity in an interval, noting that the sum of the lengths of such intervals can be
Then consider the difference between the upper and lower sums.]
made
arbi-
trarily small.
< <1
=1
[6a;-ll<a;<2
(2x
92.
93.
If
f(x)
Evaluate
{a;
a;
-h |.}
[a;]
/"17-/2
(a)'
^
Prove that
sm" X
Prove that
::
,0.5
Show that
j-
[x]
-\-
dx
rr dx
cos" X
ir
-r
for
all
exists.
f cos^ X
Ans. 9
Ans. 3
0.4872 approximately.
X dx
j
tan* x
dx
tan~' X
J'
97.
Jo
f(x) dx.
dx
J'zu'
95.
"^o
sin" X
-:z
Jo
find
where
dx
^2
a;
^ 3
_ w^
" 2V2
'
real values of
m.
x.
Inter-
chapter 6
Partial Derivatives
{x,v)
i-
We
^".u "^""l
although
function and as a variable.
Example:
If
The concept
f(x,y)
x^
2y\ then
easily extended.
is
/(3,
Thus
-1)
(3)^
2(-l)'
w = F{x,y,z)
DEPENDENT
7.
and
of a FUNCTION
a dependent variable and x and 2/ the independent
variables
m,
Ihe l^
function is called single-valued if only one value of z
corresponds to each pair (x y)
for which the function is defined. If there is
more than one value of z, the function is
multiple-valued and can be considered as a collection
of single-valued functions.
Hence
^
we
= ^i^_>y)' we
call z
^'^^l^^^
.
of definition or simply
Example:
If
having center at
(0, 0)
is
defined
is
called the
domain
constructing 3
The
x,
hemisphere of radius
and center at
= Vl-(x^ + 2/^)
(0, 0, 0).
101
[CHAP.
PARTIAL DERIVATIVES
102
NEIGHBORHOODS
where 8 > 0, is called
\x - Xo\ <8,\y- yo\ < 8
set of all points {x, y) such that
< |a: - a;o| < 8. < \y-yo\< 8 which
a rectangular 8 neighborhood of (xo.yo); the set
neighborhood of ixo,yo). Similar remarks
excludes (xo,yo) is called a rectangular deleted S
circular 8 neighe.g. {x-x,f + {y-yof < 8^ is a
The
called a closed
set.
REGIONS
point of S if there exists
point P belonging to a point set S is called an interior
point P not belonging
S.
to
belong
a deleted 8 neighborhood of P all of whose points
of P all of
neighborhood
deleted
8
a
to S is called an exterior point of S if there exists
point of S
boundary
a
called
is
S
to
whose points do not belong to S. A point P belonging
not
points
also
and
S
to
belonging
contains points
if every deleted 8 neighborhood of P
belonging to S.
joined by a path consisting of a finite number of
then S is called a connected set.
broken line segments all of whose points belong to S,
points or interior and boundary
interior
of
A region is a connected set which consists
An open region
points.
boundary
its
all
containing
points. A closed region is a region
If
set
S can be
Examples
of
a^x^b
a^x^b c^y^d
c<y<d
connected regions.
(&)
(a)
(")
c:>
X
<
Fig. 6-1
LIMITS
may not be
Let f(x,y) be defined in a deleted 8 neighborhood of (xo.yo) [i.e. f{x,y)
approaches
and
x,
approaches
x
as
y
of
f{x, y)
defined at {xo, 3/0)]. We say that I is the Umit
if for
f{x,y)-l\
lim
[or
^l
Mm
f{x,y)
approaches (a;o, 2/0)] and write
2/0 [or (X, 3/)
(a:,i^)--(xo,i/o)
X'^Xq
CHAP.
PARTIAL DERIVATIVES
6]
we can
103
number
[depending on
0<\x~xo\<8
Let
={^-
/(...)
As .
^^ = Sg-
If
-. 1
and .
/(1, 2)
and
since
/(1, 2)
is
2 [or
6.
satisfied.
0.
The
{xo,yo), in
0<\v~l\<8
< (x-XoY +
and
(v-vo)''
;
vj'
(., .) -^ (1. 2)
To prove
this
<
S^
],/(, ,)
we must
limit
would in fact be 6
3/->2
even
ii
fix,y)
Jim^ ^^Jix,y)
^^
to exist,
"
it
re-
^""'^^^ ^^^*
^"
'^ *^ ^i^^^^"* approaches
give
Sve^dSerL^v
T^'^'tt
f'^^
different values,
the limit
cannot exist (see Problem 7). This implies, as in the
case
of functions of one variable, that if a limit
exists it is unique.
The concept of one-sided limits for functions of one
variable is easily extended to
functions of more than one variable.
Example
1:
^'^'"^''^-
^/2,
a:-+0-
= ~W2
y-ti
Example
lim
2:
tan-
(2//a:)
proaches of Example
is
clear
different ap-
I'?
ITERATED LIMITS
The
iterated limits
^"d
ii^iS/(^'^)
they must be equal
| lim
ii/(^'^)
lim fix,y)
if
lim
is
/(x,
and
respectively]
limUim
a^e
fix,y)\,
[also
\U "^necessarily
equal.
denoted by
Although
j/-*;/"
lnn(
1)
-1.
Thus the
2/)
cannot
exist.
CONTINUITY
(X.
(xo
znai\r(x,y)
conditions
t(xo,yo)\
must be
<
'
"l^h^jhd
whenever
satisfied in
f i^O'yo)
is
[i-e.
f{x,y)
continuous at (xo,yo)
must be defined at
if for any positive
[CHAP.
PARTIAL DERIVATIVES
]^04
lim
1.
f{x,y)
i.e.
I,
{xo,yo)
-^
(.r,y)->(xo,o^
2.
f{xo, yo)
3.
must
exist,
i.e.
fixo,yo)
we can
If desired
Example:
If
/(a;,
= [f'
fi.,y)
t:y)
*-
= a:2y
lim
f(x, y)
.Ji^..J^^''^ =
'
y) is not
(x,y)
^^"'^^
f^'''^'
6 for
discontinuous at
a function is not continuous at a point {xo,yo), it is said to be
If, as in the above example, it is
{xo,yo) which is then called a point of discontinuity.
discontinuity that the new
possible so to redefine the value of a function at a point of
of the old
function is continuous, we say that the point is a removable discontinuity
it is conif
plane
of the xy
function. A function is said to be continuous in a region
If
Many
UNIFORM CONTINUITY
and also (xo.yo) in
In the definition of continuity of f{x,y) at (xo,yo), 8 depends on c
not on any parbut
on
only
PARTIAL DERIVATIVES
of a function of several variables with respect to one of
constant, is called
the independent variables, keeping all other independent variables
Partial derivatives
the partial derivative of the function with respect to the variable.
f.{x,y),
or
fx,
when
and
-^
fx{x,y),
-i-
dy
or
/,
needed to emphasize
it is
By
definition,
df
ei
when
,.
llZ
often indicated by
Example:
If
must be continuous
alone is not enough
fix,y)
_
"
df
'
+ 3xy\
+ 3(2)^ =
2x'
6(1)^
f^xo.yo)
fx{xo,yo)
then
18,
|^
and
/.
/(1,2)
df/dx
in
,,
At/
^^
fyixo,Vo)
respectively.
Gx"
+ Sy" and
6(1)(2)
f{x,y+Ay)-f{x,y)
/.(1, 2)
If a function /
df/By
6xy.
are
Also,
12.
df/dx, df/dy
in a region, then /
CHAP.
PARTIAL DERIVATIVES
6]
105
B_(sl\d^^
dx\dxj
If /x.
and
othervsrise
fd\_3^_
d^f
6^ =
8/3f\
,
f^^'
a^l^toj
8^f
If
f(x,y)
+ 3xy' (see
Ua.2) = 12,
2x'
In such case
In a similar
,
= ^-
fl^to
fy, are continuous, then /., = /,, and the order of differentiation
they may not be equal (see Problems 13 and 43).
Example:
is
dfsf\_
^-'
dx-
immaterial-
is
= 6x,
For example
^
'
(^)
Ly
= 6y =z
^'^
dx^dy
f
""'''
DIFFERENTIALS
Let ^x
= dx and Ay = dy
=
Az
is called
the increment in z
f{a:+Ax,y+Ay)
f{x,y).
If f[x,y)
f{x,y)
^^
where
= to^^ +
e^
and
e^
"^
V^^
'i^^^
'2^^
^'
(5)
first partial
^d^ + ^dy +
pression
a/
has continuous
a region, then
Then
e^dx
(see
derivatives in
e,dy
Problem
14)
a/
The
(4)
ex-
ft
or
aJ^^+a|^^
df
%dx +
'ldy
(5)
"'"/
If / is such that
zero as
A/
e, and ., approach
mere existence of
however, continuity of U and /
AX and Ay approach
In case f,
differentiable in
THEOREMS
on
%.
DIFFERENTIALS
If
f{Xi,X2,
..,Xn),
then
In
(6)
we
/.
'
[CHAP.
PARTIAL DERIVATIVES
106
2.
Xi,X2,
3.
If
=
dx
dz
P{x,y,z)dx
The expression
Qdy + Rdz
^^ =
if
tial
4.
+ Q{x,y)dy
P{x,y)dx
The expression
is
a constant, then df
be independent variables.
c,
Q(x,y,z)dy
if
-\-
Q dy
R dz
-X-
is
the differen-
if
an
P dx +
dQ
dR
5^' g^
an exact
is called
is called
dQ
aP
=
^
.^
and only
case
this
or briefly
R{x,y,z)dz
,
in
Pdx + Qdy
P dx
Pdx + Qdy
or briefly
In such case
Note that
0.
5^'
differential,
Proofs of Theorems 3 and 4 are best supplied by methods of later chapters (see
Chapter 10, Problems 13 and 30).
COMPOSITE FUNCTIONS
where x = g{r, s), y = h{r, s) so that
DIFFERENTIATION
Let z
f{x, y)
of
_
~
dz_
dr
In general,
^^^"
if
du
w =
dzdy
dz_
dy dr
ds
F{xi,
xi, X2,
.,Xn
ds
Xi
/i(n,
fc
dxdrk
du dXi
dxi ds
aa;2
,^.
.,
Xn
l,2,
fn{ri,
/g\
from one
rules.
HOMOGENEOUS FUNCTIONS
on
F{XXi, \X2,
Fix.y)
(8)
dXn ds
'
Example:
.,rp),
then
s,
,p
These
EULER'S THEOREM
du dxn
"
ds
Then
s.
dy ds
.,rp),
and
dzdy
dzdx^
dx ds
dud^
dud_x^
dXidrk
_
~
du
where
.,x)
dudjc^
dxidrk
drk
If in particular
dzdx
dx dr
z is a function of r
x*
2xy^
F(\x,\y)
.,
by*
(\x)*
XXn)
is
2(\x)(\yr
5{\yr
\*{x'
(see
dF
all
values of
(10)
4,
since
+ 2xy' - 5y*) =
if
F{xu X2,
.,
\*F{x,y)
Xn)
is
homo-
Problem 25)
dF
+
xi^ + X23
3a;2
-
aici
for
Xn)
.,
homogeneous of degree
if,
dF
+ - =
.
dXn
pF
,^^.
(ll)
CHAP.
PARTIAL DERIVATIVES
6]
107
IMPLICIT FUNCTIONS
In general, an equation such as F{x,y,z) =
defines one variable, say z, as a function of the other two variables x and y. Then z is
sometimes called an implicit function
ot X and y, as distinguished from a so-called explicit
function f, where z = f(x y) which
^
is such that F[x,y,f{x,y)] = 0.
Differentiation of implicit functions is not
independent variables are kept clearly in mind.
difficult
JACOBIANS
and G{u,v) are
F and
determinant defined by
If F{u,v)
dF_
dF_
B{F,G)
dU
dv
Fu
F,
d(u, v)
BG
BG
Gu
Gv
dU
dv
(7)
djF.G.H)
d{u, V, w)
is^c^alled
the Jacobian of F,
G and
Fu
Fv
F^
Gu
Gv
Gw
Mu
Mil
Hu
with respect to
v and w.
u,
Problem
3"
P"^""^^'
^""^^^^^
^ ^^d
djF.G)
d{F, G)
'
Sy
d(F, G)
F{u,v,w,x,y)
0,
Thus
du
dx
'
if
we
djF.G.H)
d{x, V, w)
d{u, y)
d{F, G)
d{F,G,H)
S{U' '> W)
'
dy
d{F,G)
d{u, v)
0,
H{u,v,w,x,y)
as functions of x and y.
In this case.
d{F,G,H)
'
we have
d{F,G)
d{u, v)
G{u,v,w,x,y)
In this case,
d{F,G)
dv
9x
d{u, v)
easily extended.
v as functions of x and y.
^{y,v)
d(u, v)
we may,
G{x,y,u,v)
0,
d{F,G)
dU
dx
dw
d{u, V, y)
'dy
dy
djF.G.H)
d(u, V, w)
33).
(see
[CHAP.
PARTIAL DERIVATIVES
108
THEOREMS
on
JACOBIANS
In the following
1.
we assume
necessary and
that
sufficient
all
equations in
F{u,v,x,y,z)
2.
,..,,,
example) is that
variables,
where
0,
d(F, G)
-q^jj^
is
m<n.
s,
^^^
then
^^^
d{u,v) d{r,s)
d{r,s)
3.
is
u =
f{x, y)
and v
g{x, y),
J
j-i,
between u and v is that
Similar results hold for n functions of n variables.
be identically zero.
form
^{u, v)
exists
J-
5(m,v)
^^^^
in
=
^
F(u,v)
are employed.
TRANSFORMATIONS
The
X
| ^
{
set of equations
(,f)\
(^0)
^J^^^j
defines,
in
general,
transformation or mapping.
{10)
a closed region
of the
uv
plane.
we can show
Then
of the
if
that
d{x, y)
lim
(li)
d{U, V)
limit as AA^y (or aAuv) approaches zero. The Jacobian on the right
of {11) is often called the Jacobian of the transformation {10).
If
u =
we
The Jacobians
(see
Problem
^
a\x, y)
45).
and
Hence
5^
o\u, V)
if
one Jacobian
is
different
from zero
in a region, so also is
the other.
ideas can be extended to transformations in three or higher dimensions.
simplicity
shall deal further with these topics in Chapter 7, where use is made of the
of vector notation and interpretation.
The above
We
CHAP.
PARTIAL DERIVATIVES
6]
109
CURVILINEAR COORDINATES
If {x, y) are the rectangular coordinates of a point in the xy
plane, we can think of
(u,v) as also specifying coordinates of the same point, since by
knowing {u,v) we can
determine {x,y) from {10). The coordinates {u,v) are called curvilinear coordinates
of
the point.
Example:
= p,
-y
= 0.
Chapter
in this case
7.
1.
f{xo +h,yo+k}
This
is
f(xo,
2/0)
k fy{xo + 6h,Vo+ek)
= ^x = x-Xo and
<
<
(12)
= Ay = y-yo.
2.
3""
'
3
is\^,
1 A a
fi^o,yo)+[^h^ + kyxo,yo)+j,(^hl
+ k^jfixo,yo) +
f{xo + h,yo + k)
+ ^(^^to +
where Rn, the remainder after n terms,
^5^j^(^'2/)
+ ^"
...
(13)
given by
is
'"*'
a
^-
(^rnyrV^s^ +
'^ai^y
o<e<i
f{xo+9h,y,+ek)
(u)
{"^Tx
^(^h^^ +
'^,^J/(-o,2/o)
=
etc.,
where we expand
Equation {13)
y - t/o. Note that
In case
limi
is
(h^
h^fxx{X(,,yo)
k~\
k^^yixo,yo)
2hkf^y{xo,yo)
all {x,y)
k^fyy{xo,yo)
{15)
= ax = x-xo and
=
w 0.
{12) is
/?
2hk^ +
= Ay
=^
an
expansion of f{x, y) in powers of a; - Xo and y-yo convergent in this region,
called the region of convergence.
This series is called a Taylor series in 2 variables.
Extensions to 3 or more variables can be made.
infinite series
[CHAP.
PARTIAL DERIVATIVES
110
Solved Problems
FUNCTIONS
1.
If
GRAPHS
and
= x^-2xy + ^v\
f{x,y)
(a)
find:
/(-2,3);
fix,y + k)-f(x,y)
-.
fi^,
(&)
fc^O.
{-2f
2(-2)(3)
(a)
/{-2, 3)
(6)
(c)
= -8 +
3(3f
vXf)^
-i-{[a;
\y)
2a;(j/
12
27
31
i i
3(2/
+ k)'] -
[x'
+ fe) +
^^^
- 2xy + 3y^]}
-2x + 6y +
3k.
Give the domain of definition for which each of the following functions are defined
and
and indicate
real,
f{x,y)
{a)
this
domain graphically.
\n{{16-x'-y'){x^
The function
is
defined
+ y'-m
all
'i)
>
0,
i.e.
<
x^
+ y' <
16
the required domain of definition. This point set consists of all points interior to the circle
origin,
of radius 4 with center at the origin and exterior to the circle of radius 2 with center at the
as in the figure. The corresponding region, shown shaded in Fig. 6-2 below, is an open region.
which
is
Fig. 6-3
Fig. 6-2
- V6 -
f{x,y)
(b)
{2x
+ Sy)
The function
3.
is
defined
6,
which
is
2x
+ 'iy + Sz =
12.
(z
= 0)
is
Trace on yz plane
12, x = Q.
= 0)
is
Trace on xz plane
12, y = Q.
{y
= 0)
is
Trace on xy plane
+ 2y -
42/
+ 3Z =
2a!
+ 3 =
6,
= 0.
The surface
points A(6, 0,
0C = 4
0),
is
BC
and
AC
A(P.n.o)
in Fig. 6-4.
B(0,
the required
Fig. 6-4
CHAP.
PARTIAL DERIVATIVES
6]
-J.2
(^)^ +
y2
^2
T?
Trace on xy plane
Trace on yz plane
IpI
= 0)
(z
(2/
As
(x
Trace on xz plane
Ill
^+ ^ =
the ellipse
is
i^
= 0.
= 0)
is
the hyperbola
^-^ =
1,
a;
= 0.
= 0)
is
the hyperbola
-^ =
=n
z=p
parallel to the
plane
a;?/
is
the ellipse
Pig. 6-5
The surface
is
6-5).
5.
M-^2
Method
1,
using definition of
limit.
exclul;
;JlViJ
Thus
1-28 + 8^
^"'
'
"'
^''"'
<r-r< f7<T-t<
"'''
<^^<i + 2s + 8'
and
Method
2,
+ 2a.
+ 2 - 5 <
a;^
r.
v,
<
<
c.'
Prove that
f{x,y)
By Problem
4,
Jim
{x"
x^
+ 2y)
+ 2y
f(x, y)
lim
a;^
1 -f
continuous at (1,2).
is
Also, /(1, 2)
5.
lim 2m
1^
+ 2(2) =
5.
6.
lim fix, y)
Determine whether
f(l,2)
v)
f(x.
'^'^^
= / ^^ "^ ^^'
to,
(a)
has a limit as
a;
and
2/
2,
(6)
is
is
continuous at (1,2).
continuous at
(1, 2).
2-a <
+ 2^, - 5 <
'
S-.2
Then
and
-^'
<
+ a,
Adding.
ie
58
i^^+^^-^i <
'''''
- 4S + 5^ <
or
Th<>
2^/
^""^^
lrr=>
'
5.
- 26 <
-56 <
that
x
'^
*^- l- <
^'
'"'
a"'
46
S^
2
+
^
+^3/-5|
<
p;I
^s
5S
i,
when-
[CHAP.
PARTIAL DERIVATIVES
112
By Problem
(a)
is
B,
-2
value at (1,2).
(6)
Umf{x,v)
follows that
it
4,
and
discontinuous at
/(1, 2)
\imf{x,y)
follows that
it
0,
/(1, 2).
'
>
(1, 2).
y^
= \ x^ + y""
[0
x^
7.
Let
a;
and
3/
in
f{x,y)
such a
x^-y''
_
lTo^^+ "
way
that y
- m^a;'
+ m'x'
a;"
,.
= 'mx
l^^ox'
i^'^'^^
{x,y)
{0,0)
xy plane).
(a line in the
_
"
x\l
,.
i'i^x^d
at (0,0).
- m")
+ m^)
Then along
this line,
1-w^
1
+ m^
(0, 0)
(i.e.
the slope
of
Another method:
Since
cannot
limjlim^X^I =
exist.
Hence
f(x, y)
lim
and
limjlimj^i =
cannot be continuous at
PARTIAL DERIVATIVES
If f{x,y) = 2x^-xy + y^,
8.
(a) df/dx,
find
-1 are not
(0, 0).
and
directly
dfldy at {xo,y^)
(6)
from the
definition.
,
(a)
a/
_
-
V-
/^
,,
J '{> y<>)
f2(a;o
lim
i,-
f{xo+h,yo)-
+ hy -
(xo
h)ya
^.^4te + 2fe'-%,
/.(*o,5/o)
lim
[2a;?
/,
lim(4^
+ yl]
ii-*"
lim/(il!l4lli(^iM
a;o(j/o
/b)
(2/0
fe)']
+ 2fc-j/o)
ft
[2a;S
Xayo
4a;-2/
[ixl- Xoyo
h_0
(*)
+ yl\ -
-i
h-fO
f(xo,yo)
""J
/C
lim (-X0
+ 21,0 +
fc)
yl]
- x +
2^.
fc-vO
Since the limits exist for all points (xo,yo), we can write
X + 2j/ which are themselves functions of x and 1/.
fx{x,y)
fx
4a; -.2/,
fy{x,y)
respect to x,
Note that formally fAxo.yo) is obtained from f(x,y) by differentiating with
diiferentiatmg
=
Similarly
l/o) is obtained by
fAxo,
x-=xo,y
putting
J/o.
and
then
constant
keeping j/
while often lucrative in practice, need
/ with respect to y, keeping x constant. This procedure,
partial derivatives are
not always yield correct results (see Problem 9). It will work if the
continuous.
CHAP.
PARTIAL DERIVATIVES
6]
/40,o)
ii^
/,(0.0)
H^
Let
(6)
a;
and
-^
+m
2/
discontinuous at (0,0).
/(x,2/) is
(ft)
/(fe.O)-/(0.0)
iio
/(0.fe)-/(0,0 )
j.^0
-^
213
= mx
xy plane. Then
in the
lim f(x, y)
lim
not continuous at
Hence f{x,y)
is
(0, 0).
Note
also
that
if
(...)^(0.0),
l7Zl2\ZcZJT
10. If
4>{x,y)
+ e-\
x^v
find
(a) ^^,
(6) ^^,
"'""
^ li ^ ^{x'y + e^')
3x'y
^^^
" l| =
x'
M
^^
^"'^
^.
_ 3V _
~ asc^ -
e'"'")
/a^N
a^VW
-0
g^i^^'y
(c) ^^^.
e^y^-y'
+ v'e^y')
^^
and .
= 0.
(d) ^^^,
(.)
3x^y
e^y'-2xy
f^
x'
6a;j/
+
+
,,,
,,(0,0).
/.(0,0)
^^,
(/)
^^^.
+ y'e^
2xye'^'
j/=(e^i'.j/)
6xy
y^e'^
= a^ = 5^ (a;' + 2x3/e^) =
+ 2xy '-(e-y") + e^y'-f(2xy)
- 2xye^y^-2xy + eTO*'2a; = 4a; Ve' + 2a; e^
_
(f\
^^^
^^,
^''
i^i'^'y
/.
^
^"'^
ay
^ ^f\ -
^V
a^
3a;*
_
=
a^Wy
2xy^e''y^
a /ac6\
3a;*
_a
"
2ye^^
,
^(a;'
2a;3/'e*
+ 2a;2/e^*)
Zx^
2xy
e='y^ -
y^
e-y^
2y
22/ew2
11.
Show
that
We
V{x,y,z)
{x^
+ y2 + ^2yu2
(0,0,0).
Then
ITT
-^ =
"a^
-i(a;*
+ 3/= + 2*)-'.2a; =
^[-^(='
+ 2/' + 2')-''T
3a!*
(a;
+ 2/^ + zT'*
-a;(a;*
+ y' + zT'"
_ 2a;* - - z*
- (x* + + ^2]
{-=c)[-Ux'
+ j/* + g*)
(a;2 + 3/2 + 22)5
(a;*
+ j/* + 2)-s/2
1/*
j/^
5/2
2x]
(a;*
+ y^ + z^"'
(-1)
'
[CHAP.
PARTIAL DERIVATIVES
114
dW _ 2y^ -x^-z^
-^ - (^r+^qrj^'
Similarly,
-J^+^+
Adding,
= :nan-|,
12. If
^
-
dz_
d^
The
'
z^yil, 1)
"51,
prove that
fxy
fyr
G =
Then
f(xo
4>(x,y)
(i)
G =
(3)
region
^^(^^^^_
1.
Zx is
continuous at
(see
(1, 1)
remark at the
and
if
/ and
fy. exist
at this point.
Define
9).
of
^^.-^
x' + y^
X
x-' + y^
\ ^ fa' + i/-)(3x-)-(a^-)(2a;) ^
we are using
end of Problem
at (1,1).
X'
-x^-y^
+ y^ + z')"^
2z'
(x'
sz^
+ {y/xYdy\xJ
3rdz\
a^z
find
1^1^^^
,
"^
ez'
d^U
Consider
+ h,yo + k) -
f(x
h,y)
-pixo,
yo
+ k) -
f{xo, yo
f(x,y)
+ k) -
<p{x,y)
(2)
,p(x,yo)
f{xo
G =
(i)
h,
/(o, yo)
+ k) -
f{x,
^{x^
j/o)
h, yo)
f{x,y)
^l^{xo,yo)
(6)
=^
H,{xo
+ e^h,y,) =
h{f.{x<,
From
and
(7)
(8)
G =
G =
fe
-^
and
e,h,yo
+ k) -
and
to (5)
(6),
f.(xo
< 1,
< 1,
tf^
ff^
<
<
9,
<
<1
94 < 1
ffj
fy4xo
fe
-*
+ e^h,yo+e^k) =
in (9)
we
f^y(xo
have, since
/x,
fyx(Xa,yo)
and
e^h,ya
/,x
+ e^k)
first partial
(a;o, J/o),
fxy{Xo,Vo)
Problem
43.
DIFFERENTIALS
14.
<
e,h,yo))
we have
<
<
+ e^h,yo + e^k)
hkf:^{xo + e^h,yo + 9^k)
hkfy.{xo
as required.
derivatives in a region
of the xy plane.
Prove that
A/
where a and
=
e^
f{x
+ Ax, y + Ay) -
f{x,y)
U^x +
fyAy
c.ax
..Ay
A/
(i),
we have
(9)
Letting
=
=
(see
Page
61),
we have
2/
9,
<
^^'^^-
PARTIAL DERIVATIVES
6]
Since,
by hypothesis,
and
/,
are continuous,
e^ -* 0, e^ -^
Thus
Ao; -
as
follows that
it
= Mx,y) +
and
II5
fy{x,y
15.
e,
A/
call
U^z =
fy(x,y)
A2/ -> 0.
Defining Ax
We
+ e^^y)
= x^jz-S^/,
-0.01, A2/ = 0.02.
f{x,y)
y~Z, Az =
find
(a)
(d)
How
e,dx
e,dy.
(b) d^.
(c) Determine Az and dz if
x=4
might you determine /(5.12, 6.85) without direct
A^,
computation ?
Solution:
(a)
Az
{Axf Ay
'
'
(A)
is
^*^
(d)
= 2xyAx +
'^^
Another method:
(c)
Jb)
Az
=
=
/(x
+ Ax,
d^
and
AT^d^ltZL:^Zr'
that
/(4
{(4)=(3)
= 2xydx + (x^-3)dy =
2x2/
dx
(a;^
+ Ax =
a:
- 3) dj/
3)
+ Ay =
2xy dx
(x'
direct?omUtio7fll9.oSl^^
- 3) dy =
154
2(5)(7)(0.12)
+ 5.1 =
0.02
and
5.12
(5^
We
6.85.
f(x,y)
dz,
- 3)(-0.15) =
approximately.
159.1
Thus
z, i.e. dz.
(x^-z)dy
(4^-3)(0.02)
^" ^PP-^-t^I^'
"'
2xy dx
2(4)(3)(-0.01)
/(r.
(x^-3)A2/
/(x,j/)
3(3.02)}
dz
Wemustfind
the differential of
= grf^ + gdj/ =
+ A2/) -
2/
{(3.99)^(3.02)
is
= d.
can accomplish
i.e.
+ dz.
5.1.
16. (a)
(a)
Method
1:
S
^''^"
Method
Method
If*^"^
^-"
^dy
'"(;
+
(2xey"-yey")dx
xe^'^
dy
2:
=
=
x^die"")
ey'-d{x')
^dy~ydx
^.^y,.(
\^
x^e"d{y/x)
2xe''da;
2x6^'-
(2xey''
dx
ye^") dx
1:
Suppose that
Then
'^^
dU
(6)
= '-"(-)
(^x^y
- 2y') dx +
(,)
(x^
- ixy + &y^) dy =
= 3.V -
2y^
(.)
fj
= ^dx + ^di,"'
dx
dy
dd,
,3
4,^
6^.
a;e'-<
PARTIAL DERIVATIVES
116
From
=
is
x^
F'(y)
x'
is
4xy
2xy'
F{y)
4:xy
x'y
from which
Gy'
F'(y)
2xy^ + 2y^ + c,
x^y
we have
(1),
where F{y)
[CHAP.
where
yields
(2)
6y\
F(y)
i.e.
2y^
an arbitrary constant.
c is
Note that by Theorem 3, Page 106, the existence of such a function is guaranteed, since if
- 22/' and Q = x' - Axy + 6y\ then dP/dy = Bx^ - iy = SQ/dx identically. If
dP/dy = bQIbx this function would not exist and the given expression would not be an exact
P =
Zx^y
differential.
Method
2:
{Zx^y
- 23/") dx +
('
- 4*2/ +
dx
(Sx^'y
x' dy)
2xy^ + 2y' + c.
x'y
is
=
=
=
dy
62/'')
This method, called the grouping method, is based on one's ability to recognize exact differential
combinations and is less direct than Method 1. Naturally, before attempting to apply any method,
one should determine whether the given expression is an exact differential by using Theorem 3,
Page 106. See last paragraph of Method 1.
DIFFERENTIATION
17.
Let
of
and x
f{x, y)
COMPOSITE FUNCTIONS
= 4>{t), y = f{t) where /,
smce as At
= e^
dt
_
"
dz_
18. If
_
"
dz
dt
A-o
At
we have Ax
COS
t,
By At
-^ 0,
Ay
sin
(v'e=^y^){- t sin t
= f{x, y) where x =
.,93 _ dzdx_dzdy
'dxdu
^"''du ~
dydu'
(a)
</>(tt,
From Problem
dz
du
(b)
20.
The
14,
Az_
(h\
a"-*oAm
Prove that
dz
dz
3a;
dy
t)
dx
even
if
Ay
dx
= xdu +
Xvdv
+ Xwdw
dz
dy
dy dt
-jr
Tr/2.
(0)(1)
and then
x and
1/
^1
'Am
'
sin
t).
-,rV8.
differentiate.
(2)
Au j
letting Ai;
9z dx
dz dy
dxdu
dydu
-^ 0.
dx
dy
{2xye^y'){t cos
by replacing Au by Av and
-dx+^dy
5z
dx dt
53 a^
dydv'
/, 0, ^t, we have
32 to
dxdv
dy Au
At J
prove that
^{u, v),
^^
^'
-^ -tt
'''
"'"
e*
differentiability of
^^
Au^o^dxAu
j
dz
COS
^'dv
A.X
(,rV4)(-,r/2)
and y =
v)
assuming the
= 0,y^7r/2. Then ^\
^
0,
"
^Jl\.
At
compute dzldt at
t,
19. If
*'
e^^
e^ -^ 0,
dz\
Att = TT/2,
^+
i^^_|_^4i
[flccAf
=
|5^
+ |5^
ay at
"X dt
dzd^
dy dt
At-.o
-^ 0,
dzdx
dx dt
we have
14,
./r
<^,
^^^^
dt/
Then
= y^du +
yvdv
+ ywdw
Prove
CH-^P-
PARTIAL DERIVATIVES
^]
Thus
z.dx
=
=
zydy
+ 2y) du + (z.a;. +
+ Zvdv + Zy,dw =
(z.x.
Zudu
r = x^-xy + y\
21. If
3T
may
]L
(a)
psm<j>,
l^J^^d^a^ =
dT
dT
dx_
aa;
30
also be
dr
dp
^JL
^ dU
dx dr
dT dy
find
dT/dp,
(a)
iSx^-y){cos^)
_
~
30
32/
+ 2s,
3r^
dy_
ZyyJ)
dw
(b)
dT/d<t>.
dT dy
dy dr
(^.a;^
dz
^^^^
~ 2/)(- P
1/
{Sy--x){sm^)
sin 0)
^-^^^^y'^ where
^^'
pcosi>,
dTdx
d^
z,j/.)
19.
_
~
30
This
II7
(Sj/^
- x){p
cos 0)
in T.
- 4r-2s3,
= 2r2-3s^
find
(a)
flCZ/flr,
dz^
dz dr
"'
(6)
dx 3s
dUdxdUdydUdz
ds
32/
z cos
23. If
ds
dz 3s
f)(-?)}'^'+{( f)}<-)+(='nf)M..
22/?
= pco^^,y =
6s'z
2/
show that
p sin ^,
(^^ + (^^X
(^^X+
=.
(^^^
=
=
Vp
V^
Dividing both sides of
(2)
Then from
we have
by
p,
V.Xp
V.x^
+
+
'^p
24.
Show
that z
Let
x^'y
and
(5),
= u. Then
a:
ycos0 + y, sin0
Vyy^
y. (-p sin
0)
yj(p cos 0)
~ ^i
is
Vy sin
0)2
sin
dz du
2*;^,
(2)
Vy cos
(- y^ sin
differentiable,
(3)
satisfies
y,,
cos 0)^
xidz/dx)
= f(u). Thus
/'(k)
(Vx cos ^
where
fix^y),
dz
Then
-^Vl
we have
~^<l>
(1)
y,2/p
...
22/
1^
"
dz
2;.^
and
dz
so
dii
a;|i
3a;
22/^.
32/
V^
2y{dz/dy).
V^
[CHAP.
PARTIAL DERIVATIVES
118
Another method:
We
have
dz
f'{x'y)d{x'y)
Also,
dz
= ^dx +
Then
dx
-r-
dy
x" dy).
dy.
2xyf'{x'y),
f'{x'y)(2xydx
x^f'ix'y).
dz
a;r
25. If
for
all
Let Xx
= u,
\y
= v. Then
= XTix.y)
F(u,v)
dF
SF du
8F
d\
du dX
dF
26. If
F(x,y)
in (2), so that
x'^y^sin-^y/x,
F(Xx,Xy)
Since
27.
du
-ri
dv
Then
dv
=1
SF
_
"
dv
dv dx
.^+yf
du
Letting X
(1)
u=
x,
= y, we
show that
pX^-^F
have
x{dF/dx)
x{dF/dx)
(2)
y(dF/dy)
y{dF/dy)
pF.
6F.
(\x)*(Xy)^ sin''
= 6.
It
Prove that Y = /( + at) + g{x - at) satisfies d^Y/dt^ = a\dW/dx^), where / and g
are assumed to be at least twice differentiable and a is any constant.
Let u
dY
By
d^Y
^
'
Si^
x-\- at,
dY du
dY
at
Y =
so that
dv
,,,
f{u)
,,
dYt^
dt
_
~
dYt dv
aFt du
du dt
dv dt
-^
f"{u)
{af'(u)
aV"W
2 f",
+
^
d^Y
(^)
dYr
dYx du
dYx dv
,,,,
Then from
(i)
and
(2),
S'^F/at'
a^d-'YIdx').
d^f/du'^,
ag'(v)} (a)
f'(u)
d/Zdw,
g"(v)
,,
p"(^)
\^
-^ {af'(u)
mi
fl''(-D)
,,,
d^g/dv^,
a^9"{v)
dx^
= /"W +
if
dY du ,dYdv
dY
_
~
+ g{v). Then
//
\i
dgfdv,
,,
we have
ag'iv)} (-a)
CHAP.
6]
28. If
PARTIAL DERIVATIVES
= 2r-s and
Solving
Then
dr/dx
2r
- s,
+ 2s,
as/aa;
2/5,
find
2s
8U
_a_
^2
ar;
(2^+
^
25 V
U
t,
2acr_iac7
5 dr
5 as ar
(1)
and
dr ds
Solving
(5)
dx
and
dt
dU
Then
n
5 ds
J dy
a''t/Y2
5 ds'
2t
Zy'
5x*
2x
32/
J\5
dUdy
t.
dy dt
t\
(implicit) functions of
2x(dx/dt)
(4)
dt
-2t
(Zx'y}
(
15x*y'
- 2x
a function of x and
aF
(')
F{x,y,u,v)
aF
5x*
2x
2t
5x*
2x
3y'
(2)
and y
in a region
of the xy
sf
aF a^
a^ + "ai a"^
0,
find
(a)
du/dx,
and
(2)
(6) flM/ai/,
(c) Sv/Sa;,
'
of x and y,
Ma:dx
+ Uydy
(i)
dv
v^dx
(d)
S-y/fli/.
- 2x
\l5xV - 2x
u and v as
Using the subscript notation, we have
ISxV'
yiOxH - 2x_
(x'
(1)
2f
^^^
"
y.
Zy'(dy/dt)
dw.
+ T
dy "
az
a^+a;a^ -
(S)
- dx
dx
we have
and G(x,y,u,v)
dx
i/,
Then
diz/dt,
Zy'--2t
15a;V -2x
ac/d X
ax d t
dy/dt
+ y^ =
x^
Also, since
dr
ds'
(^)
dt
and y as
(2) define x
5x\dx/dt)
(3)
31. If
_ lacA as
a_/' 2 at/
ds\5
J dy
^2 dHJ_ _
5 as"
Equations
we have
Since z
Hence we have
2/5.
i3f7\ ar
5 ds
+ y)/5,
s.
= {2y - x)/5.
-.IE.- o!E
dr'
assuming
_
~
ds
2dHl_l i^C^yi^ +
_^
t;
(2x
as/S^/
ds dx
dr\hdr
dy
s:
1/5,
du
dr^
dr dx
d_fdU
\dx
dy dx
dr/dy
du
ax
_a^
and
f or r
-1/5,
in
dydx
119
v^dy.
[CHAP.
PARTIAL DERIVATIVES
120
Substituting
and
(5)
=
=
dF
dG
(5)
(6)
Since x and
2/
and
(i) in (1)
(2)
yields
(F.
(G.
Solving
and
(r)
(8)
= -Fx
=
G
|GM. + G.v,
Ux
(c)
ax
-Fy
-Gy
ly
F.
d{y,v)
du
1-12V
dx
8uv
aw
We
_ -2-4-u
1 _ g^t^
mnemonic
(see also
Problem
du
dv
dx
dx
dv
dx
dy
33,
,.
If
F{u, V, w, X, y)
,,.
^
'
dy
dx
9{u, v)
-Fy
-Gy
a(F,G)
F,
S(F.G)
G.
d{u, v)
F
Gu
J,
or
^\J
,s
32/
4m
dv/dx,
(6)
to x, considering
3m
Sf
dv
(c)
du/dy,
{d) dvfdy.
u and v as functions
of
du
- Sm'U
Suv
F =
G{u, v, w, x, y)
dw
the Jacobian of
u,v
u'
F.
G.
Fu
G.
= 0.
- v -Sx- y =
G = u-2v^ - x + 2y =
0,
-1
-3
-1
4-u
F.
2m
-1
F.
G.
41)
obtained.
Similarly the other partial derivatives are
0,
i
is
S(u,y)
d(u, v)
0.
d(F,G)
G.
d(F, G)
- 8mi; ^
d(u, x)
we have
d(x, v)
dx
a{F,G)
33).
d{F, G)
<?s
dv
'
provided
Fs
2m-3
SM-y'
du
dy
2m^
dy
-^ -
3(F, G)
Fu Fv
/
denoted by -.
d{u, v)
Gu Gv
and is supposed = 0.
ij
'
iS^
^^'
Hence we
-F
-G
F.
G.
B(u, v)
Solvmg,
=
=
F
G.
HF, G)
to devise
'
Vy
(d)
Solving,
are zero.
(S)
F.
G.
dV
dx
ejF.G)
Uj/
d{u, v)
F.
G.
32.
Us
Go
dJF.G)
^'
(6)
F.
G.
Gu
Gu Ms
d(F, G)
F.
G.
Fu
G.
du
li,
and
gives
-F. F.
-G. G.
(a)
G.Vy)dy
Fu Ms
(S)
=
-
F.Vy)dy
in (5)
obtain
^'^^
0,
H(u, v, w,
x, y)
0,
find
1-12V
1
8mi;
0.
Then by
^^^^-
PARTIAL DERIVATIVES
^1
j21
'^^"'- ' ^--bles in terms of
theremr4T*i:3\Llat:sl^^^
S.S., .e
know
that
wo.a .no.
^t:z^::^^'^;tzx^ :sst^:s^::
we
i.e.
serves to indicate
is
(a)
Vy,
we have
F
dv
T
^y
Equations
{!), (2)
and
31.
H^ H, H
d{F,G,H)
S{u, y, w)
F.
G
F
G.
F
G
diF,G,H)
/?
H.
H^
d(u, V,
w)
(S)
F
G
F
Gy
(?
31.
s^sriaS: r-^:r=:tj:^^
^^^^
d(F,G,H)
d(F,G ,H)
du
a(F,G,H)
d{F,G,H)
^(''y'^)
34. If
z^~xz-y =
-^
dy
prove that
0,
d{w,x,v)
= _
dx
J^+^
{3z^
xf
~ ^to ~
"
and
U)
^^^^-1
Differentiating
^^^
a.a.
The
'''
(2)
_
-
and
-1
/-.
az
(3?^(6z--iJ
(1),
we
^
dx
""'^
Sz"
(.)
= "
-X
_^
we have
~j^^^
1
'''
find
6zrz/(3z^-a;)l
""^'^^
the
=
to
_l
-0^^
y and using
(2).
continuously
^
differentiate in some
' '''
regiorJ^'pL:eV.t^'^'''^'
necessary and sufficient condition
* ^
that there exists a
!
hI
i\u
honal relation between . and of the
.
form ,(., .) =
is the vanishin/of
Jacobiat
'^-
di^
Te
Identically.
[CHAP.
PARTIAL DERIVATIVES
122
We
Necessity.
^(^''")
To do
identically.
if
d{x, y)
= ^^du +
= (<p,u. +
d^
Then
0Mx
(1)
Now
<f.dv
(<f>Uy
it
(2)
from
follows
Uydy)
4>,i^,)dy
-P.v.
,pAu.dx
<f,,v,)dx
_
-
d(u,v)
is
<t>.vy
^ aKji ^
that
(2)
.- .,
T
I.have to prove that if the Jacobian ^^^-^
0.
v)
0(m,
tional relation between u and v, i.e.
Vydy)
functional relation.
they were, there would be no
if
and
(1)
+ <f.Av.dx +
=
f>.y-y
We
Sufficiency.
exists,
note that
we
this,
,(u,v)
Uy
identically.
d(x, y)
Vy
a func-
f.
identically zero
is
and
From
these
{S)
du
(-4)
dv
From
we have
respectively,
(3),
u.Fu
and u,Fy
+ m, =
or
= v.F.du + {vA-uJu.) +
dv
(e)
d(u, v)
But by hypothesis
Ux
Uy
Vx
Vy
^(^^
(5)
UxVy
UyVx
(v,Fy
+
+
Uy)
dy
Vy)
dy
U) becomes
this,
UxVy
= v.Fudu +
Vy}dy
Using
-Uylux.
(u.Fy
UyVx
dy.
Ux
identically,
so that
36. {a) It
(b)
u=^
4>(u,
^^
Are u and
t;
d{u, v)
and v
= tan^'x +
functionally related?
Ux
Uy
Vx
Vy
(1
r/.i'
find
v)
-^^^y
+ x'
{1-xyY
if
xy^l.
d(x,y)
By ProWn.
other.
9(2^,
tan-^y,
- xyY
+ x'
+ y'
1
1
becomes
i))
1+j/'
(a)
(S)
Then substituting
tan (i;-
tan-, 2/),
this in
^ =
(x
= tan- x +
_
-
tan
i
^;
+ y)ni-xy)
'<f
tan
tan
;! ^l' ^ :
'j/
we
find
tan
t;
and simplifying, we
tan
^an^+ ytanv
j/
find
u = tan^.
i"'*Tl
2/
an
CHAP.
PARTIAL DERIVATIVES
6]
123
>.2_
X = u-v + '^'
w, 2'
y =
u^
= '-^^-^^
and z = u^
u' + v,
^/^' ^' ^^
v. evaluate the Jacobian
Jac
V.,:~
and
(b) explain the significance of the
d{u,v,w)
e non-vanishinff
non-vanishing of this .Tamhinn
Jacobian
37. (a) If
^
s(x, y, z)
(a)
d{u, V,
w)
^u
fl?i>
Xu
yu
yv
y-K
2m
-1
2v
Zu
Zx,
Zv,
3m^
6tyM^
+ 2u + 6uH + 2w
(6)
%^
if
ji
tion
y^u-v.
%
= u + v,
The region
shown shaded
9?
"
respectively^"
(b)
Compute
and %'.
6,
is
a{x, y)
^g^.
<'{u,v)
(c)
a;
-^ =
and
mapped under
Compare the
2/
0.
(a)
De-
the transforma-
result of (6)
v
) with
in Fig. 6-6(a)
*'"'
/
.^y
/
/
y=
{a)
xy plane
(6)
uv plane
Fig. 6-6
I.e.
inthe^u^rilw
m
the uv plane.
^ *"'""'''
Tl^
In
manner, y =
(u
like
2l7L''^Te-eir
a(a;,
(6)
(c)
y)
d{u, v)
dx
du
dx
dv
Sy
du
dv
dy^
+ v)-(u~v) = 2
becomes r* =
1;
''''
(u-v)^e
or ^
or
(u + v) ~{u + v)
iu-v) -(u~v)
-1
r
Sons VtTi:
'
p'l^L!'
39.
[CHAP.
PARTIAL DERIVATIVES
124
region
and 1/-0,
xy plane is bounded by x^ + y' = a?, x^ + y^ = h\ x^Q
under the
mapped
is
which
%
(a) Determine the region %' into
what
Discuss
(b)
0g^<2,r.
x = pcos</., y = psin^, where p>0,
in the
0<a<&.
where
transformation
happens when a =
Compute
(c)
0.
-g^.
Compute
(d)
^^^.
%'
= a}
*
O'
''
(6)
Fig. 6-7
(a)
The region
x'^
+ y^ =
'^l
x^
a" (dashed),
y''
and p =
becomes
a,
a^ x^b
= 0,
^p^
is
x'
Also,
respectively.
is
bounded by x=
(dotted),
(dotted
and dashed),
6' (heavy).
+ y' =
x^Q,
a?
and
a;^
a^ySb
+ j/^ =
6^
becomes
become
- jr/2,
p^
= a^
and
=p^
p^
= 6^
y-U,
b;
6.
shown shaded
(6)
S W2
6,
b (bounded by 3 sides)
= is mapped into
x
=
point
the
0,y
that
is
this
for
reason
The
while %' remains a rectangle.
which is somepoint
this
at
one
one
to
= an indeterminate and the transformation is not
p = 0,
If
the region
d{x, y)
dp
(p
cos 0)
4- (P c^s
~
dp
(p
sin 0)
i00
(c)
d(p, 0)
=
{d)
p(cosV
From Problem
45(&)
+
we
*)
cos
sin 0)
sin
(p
sinV)
3(p, 0) 3(a;,
p cos
have, letting
d{x,y) d{p, 0)
u-p,
<p,
a(p,0)
^
so that, using
(c).
3/)
sin
00
d{x, y)
it is
clear
1
p
why
(i.e.
a;
= 0,
2/
= 0)
a singular point.
Prove the
first
Let F(t)
variable,
^^^^-
PARTIAL DERIVATIVES
^]
If
x,
+ ht,y =
+ kt,
y,
= Md./dt)+Mdym)
F'(t)
where
<
<
where
<
<
Thus
1,
fixo
(1)
==
then F(t)
hf.
and F'ie)
kf,
by Problem
17,
ok)
so that
f(x,y),
125
becomes
+ h,vo + k)-
fix,, y)
hf^ (^
+ eh,y, + ek) +
+ eh,
h /, (a;
y,
F(t)
andifi =
(2)
-.a.
more^;::^Sit^^tS^,::.iil---U^
41.
+ ek)
as required.
F(0)
+ ElMJl +
2!
^
P'(0)t
l,
Prom Problem
...
^'"'(0)
,.
*
ti!
F<'^"{e) ,_,
+ \n +
i
40,
Uj^
+ kMx.,y,) =
where we have used the symbolic operator
notation. Similarly,
i^'(O)
hf4x.,y,)
l)\
o.
be..
for functions
0<e<t
(1)
+ i).
fcf>(.o,^)
^
from which
^"(0)
'^V..(x,2/)
2;ifc/.,(x,2/)
(h^
fcV.(c.o,2/o)
that for
where
42.
< < 1.
<P
Expand
Substituting these in
.'y
4-
3. - 2,
/.
. 2.y,
/.
-2)
..
+ 3,
= 2y,
/.
= 2x,
- y
/. . 0,
where
/. = 0.
= 1, y,^ -2 Then
/. . 2, /,, = , /,, = o
x,
Thus
/.,(l,-2)
/,(!,
-2)
= 2,
=
/,(!,
-2)
Taylor's theorem,
fi^,y)
where R^
is
/(I,
./.(I, -2)
is
positive integers n,
x^y
By
y,)
(2).
/(., .)
all
+ k^"^ fix,
+ By-2 = -^0~4i.-l) +
,iy
we
+ 2)-2(x-ir +
find
2i.-l)iy
+ 2) + i,-iriy + 2)
[CHAP.
PARTIAL DERIVATIVES
126
MISCELLANEOUS PROBLEMS
43.
Let
f{x,y)
\^ +2/ /
Compute
If
(a)/x(0,0),
y)
(a;,
(a;,!/)
.0
(0,0)
(c)/.40,0),
(6)/.(0,0),
/(0,0),
(d)
(/)/.x(0,0).
(e) /x(0,0),
(0, 0),
.y_^^M!_^ +
af...A^-.=_M
,(5^
Then
li^MO^i^L^AM =
(e)
/.(0,0)
(/)
/.^(O.O)
= lim/.(M)-MO,0)
Note that
44.
Show
ew
5-
/.
/.
,.^^^1
We
dW
^
becomes
-r-s-
IdV ^
-
-1
at (0,0).
li^njf
"
^r
cos ^,
= psm^
the equation
^^^
dW _
1
2
aT2
"
have
(-'^
Differentiate
tions of X
a^
'
p cos 0,
"^
a;^
1/
a0 ax
flx
with respect to
p sin
dp ay
sy
'
x,
B4,
remembering that
dy
p
and
are func-
and y
-psin0g+
cos0|^,
PCOS0U+
sin^ll
Solving simultaneously,
dp
-^
-psin0g+
_
~ _
30
V"
3x
^
cos d>,
ax
sin0
ig)
Then
cos^l^,
pcos0|j+ sin^l^
Solving simultaneously,
02/
a?/
Then from
(1)
dv
and
(2),
3V
sin
ay
,..
ay
-^
ay
cos0 dv
CHAP.
PARTIAL DERIVATIVES
6]
127
Hence
dx\axj
3x'
^ /lo ^
"^
dp\ax/di
^^
d4,\ax/di
SV\dp
sin
sirn6
aF\a0
d<j>JSx
ap
sin
which simplifies to
^'^
Pn^^g!Z
dx'
and
(7)
p''
^-
sinv
2 si"
^p'
we
()
cos
3(a;,
a;
= /(%,?;) and
2/
-a^y
-
2 sin
_|_
d<f,
= g{u,v), where m =
sm0
fXsmVg ay
d'v
apa0
cos
o
p
^"^
find, as required,
dV
ap a0
ay
cos
p'
cos^i
Stf,
2sin0cos0 d'V
^0
dx'
45. (a) If
<b
dp
sin0cos0 ay
^p^
Similarly,
Adding
ay
+
op
""
gin^A gn^
ap
cos^0
ay
ap
"^
d^v
lay
dy
ap^
p dp
p^
+
1
,Kr,s),
a^^
cos^ a^y
O V
p^
p2 30^
d{r, s)
y) d(u, v)
d(u,v) d{r,s)
(b)
(0)
(a)
di^>y)diu,v)
ProvP that
Frove
^^^ ^^^-^
_
_
XvVr
XuUs
+ yvVr
y,cUs
+ yvVs
XuUr+
d{x, y)
S{r, s)
y,Ur
y-
V'
Xu
d(x,y)
^
^^
+
provided
Xr,
Mr
Ms
J/
Vr
Vs
XvV,
a(a:,
2/
0,
y) d(u, v)
d{u,v) d(r,s)
I
= x,s = y
Place r
Then
^(^>
i/)
We
=
a(x, y)
a;j/_plane
V-
46.
Show
Let u
<^)
that F{xy,z~2x)
y{dz/dy)
What
2x.
= xy,
satisfies
= z- 2x. Then
F{u, v)
dF = F.du + F.dv
and
= F.(xdy +
ydx)
+ F.(dz~2dx) =
TV.
Then'^^^l^'^J!:^
substituting
in the
4,(x y)
of each other.
(i),
we
we have
find
(yF^
Fr,z^
2)
dx
(xF.
F^zy) dy
yF,
+ F.z.-2 =
(3)
xFu
F,Zy
dz
a^^
a^y
"*"
and v
(?)
""
a^y
ct>ir,s)
sin
z.dx
^ ^
z,dy
-r zyu.y.
[CHAP.
PARTIAL DERIVATIVES
128
The result
that Fr - 0.
xz.
we
that F{u,v)
we assume
is
- yzy =
2x
upon
Supplementary Problems
FUNCTIONS
47.
If
48
If
= x'-yz + 3xy,
g(x,y,z)
-1,
(a)
-3),
2x
W
,^.
g^g^Tpg).
(6)
-i,
(a) /(I,
find
11
ii.\
{a)
Ans.
49.
2^,
/(^,^)
Ans.
GRAPHS
and
^^'
(6)
^^''''
'i~
/(^
+ 2/,x2/).
+ 2y + xy
t-x'y-xy^
find
(a) f?(l,-2,2),
{b)
+ ^'
g{x + l,y-l,z'),
(6)
(c)
g(xy,xz,x + y).
(c)
Ans.
x^
x^
(a)
and
domain graphically.
() /(^.J/)
real,
+ y^-l
+ y"-
\,
(^)
'
(b)
/(^'^)
+y >
=
0,
ln{x
(c)
+ y),
(c)
2x-y ^
x+
= sin->(^^^.
f(x,y)
,
't
50.
(a)
(6)
What
Ans.
51.
(a)
is
+z ^
1,
x'
+ y^ + z^ <
and x
f{x,y,z)
+y+z ^
^ J^^.^^^2 _
1,
x^
+ y' + z^ >
1.
the following,
Sketch and name the surface in 3 dimensional space represented by each of
(a)
(6)
(c)
+ 2z = 12,
42 = x^ + y\
z = x'- Ay\
(d)
3x
(e)
if)
+ z' = y\
x' + y' + z' = 16,
x' - 4y' - 4.z' = 36,
x'
iff)
+ y' = 2y,
= x + y,
y' = 4:z,
+ y^ + z^-'ix +
x^
(h) z
(i)
(i)
x''
hyperbolic paraboloid,
Ans. (a) plane, (6) paraboloid of revolution,
(g) right circular cylinder,
(e) sphere,
(/) hyperboloid of two sheets,
cylinder, (j) sphere, center at (2, -3, -1) and radius 4.
(c)
52.
53.
54.
is
x^
y^
z''
1,
x^
y"-
z^;
(b)
(x, y, z)
x^
+ y' =
a'
and
x'
+ z'
=^
Gy
+ 2z-2 =
(d)
{h)
0.
a\ where a
is
a constant.
such that:
+ y" <
<
+ y.
The level curves for a function z = f{x, y) are curves in the xy plane defined by
They provide a way of representing the function graphically.
is any constant.
f{x, y)
c,
where
CHAP.
PARTIAL DERIVATIVES
6]
129
55.
lim (3^-2y)
(.)
and
14
(6)
Jim^^^
^^
(x,yy^
(2, 1)
y-+ -1
56.
If
lim/OK,j/)
(a)
hm{fix,y)
lim^^
(6)
(^)
rd)
lim ^-5ilL(2il!l
+ x~2y
il"? 4
-*
As.
59.
^' S'"
(6)
Formulate a
/A
1-
2a;
-J/
sin"' toy
-2tan
''
(c)
8\/2,
(d) 0,
(e)
0,
(/)
(a) 3,
(6)
62.
() -^
+ y^;
Ans.
(a)
(xo, yo).
continuous,
Using the
definition,
^^'
oiT^tts^'g
^^
(6)
(6)
^) -. (0, 0, 0)
(O, 0).
exist?
(c) (x^
discontinuous,
?"'''"" ^ ^'"'"' ^^
sin
if
(X, y)
if (.,
(0, 0),
?/)
continuous
(c)
y^)
6)
n variables.
61.
(a;, j/,
-2^
'(3a;j/
(h) 1/3
(g) 0,
Does
as
^ + ^"1
j;^
(^)
60.
lim
prove that:
P
lim e-'^<.->^
()
2^-^^
{x,y) -^ {xo,yo),
>!">
definition.
exist.
s_.o
(a) 4,
as
by using the
is
- 3. + 4) =
57.
(xy
xy
+ 6x
is
continuous at
uniformly continuous
*'
(a) (1,2),
in
(6)
(x,,y,).
PARTIAL DERIVATIVES
64.
If fix,y)
1^,
by differentiation
65.
If
Ans.
rules.
(^'
f{x,y)
- 2/)/(a' +
2/)
|0
Aws.
66.
(a)
1,
Jim^^^ f4x,y)
,^
-2,
for {x,y)
for (x, 2/)
3f/Sy at (2,-1)
(6)
from the
definition
-4
(0,0)
(0,0)'
..,,,,,
(a) /, (0,0),
,,,
(6)
find
/, (0,0).
why
this limit
Ans. (a)l(.
V3),
^^'
2z
(a)
(6)
(6)
Investigate
(if it
and
ff''rr?L'^/n
m*
If (x,y)
(0,0).
(6)
i(2.
3^3).
(c)|(.^-2),
(d)
|(,V3
3),
(e)
1(2.^3
1),
(/)
1)
yOz/8y)
1(2.^/3
(0,0)
69.
Verify that /
mdicatmg
70.
Show
that
ln{(a5-a)^
(2,-6)^}
satisfies
aVfla;^
aVSj/^
^t^^xy
and
except at
(c)
cosh
osn
(a, 6).
(2/
^2/
cos x),
a;)
71.
[CHAP.
PARTIAL DERIVATIVES
130
Show
a;
that z
x cos
(ylx)
tan {ylx)
x^z^x
satisfies
2a;j/2x
J/^Zss
= 0.
(a;
a;
+ z"N"
+ 2/
J/
2/
DIFFERENTIALS
73.
= x^-xy\- 3y\
If z
why Az and
Ans.
74.
-11.658,
(a)
Compute
compute
and
Az
(a)
^^(3.8)'
-12.3,
Az
(c)
(c)
= -66,
a;
Explain
3/
if
j/
= -123.
dz
using differentials.
approximately,
2(2.1)'
(6) dz where
Find Az and dz
Ans. 2.01
75.
76.
Find dF and dG
xy' In
(j//a;).
Arts,
(a)
(d)
77.
- 'iy^) dx +
{Sx'y
V-
(6)
(81/
(c)
{y^ In
Prove that
F(x,y)
(o)
if
(dy)/(l
(6xy
y^)
dx
DIFFERENTIATION
of
U{x,y,z)
(a) If
(b)
H(a;,2/)
If
Ans.
>
(a)
If
F(a;,2/)
(d) d^F/dv^,
80.
If
Z7
81.
If
a;
12t
where m
that
if
a;
V sin a and y
p cos ^,
dy
bx
Use Problem 82
to
2/
Bu
8v
IT-
dy
if so, find
the function.
= *'-* + !.
= t' + 3t,
,
+
4.,
3^u
aa;"
32/"
Ans.
?)
cos
(dVldyf
36"',
dC//dt at
find
= 0.
find dH/dt.
find
a,
aF/aw,
(o)
-14,
(a) 7, (b)
(c)
x(dU/ax)
(b)
aF/av,
+ y(dU/dy) =
where a
is
{dV/dur
{BV/dvy
(c)
-49
217.
the equations
dv
^rdx
= u + 2v,
aw
become
av
+ T-i
3a;2/
Jcos{Sx-y)
sin a
P sin
42/'
suitable restrictions on F,
du
^
<;-2/^
2t',
= 2u-Zv,
= 2, v = 1.
(dV/dxr
Show
2/
a;z'
+ 9a;"-6tM-6xH + 18\
+ y)/{y -2x),
(c)
e^vTl
d^F/du dv,
cos a
not exact,
+ 21/ =
a;'
x^F(y/x),
= M
F(a;,2/)
dz
a;
sin(3x-2/),
(2x
(e)
3a;z^
COMPOSITE FUNCTIONS
= 2 sin t,
2a;" - 2/z + a;z^
/36t=2/
^
/l^
(b)
24,
(c)
(2xe
83.
- Sx^yz,
- x^) dy
3a;) dy +
31/) dx + (122/^
(e) (z'
Ans. (a) x'y'^ + y sin 3a; + c, (b)
82.
Sxy^z"
(16a;3/z'
79.
G(x,y,z)
(b)
Determine whether each of the following are exact differentials of a function and
(a) (2x2/" + 3y cos 3a;) da; + (2x^y + sin 3a;) dy
(b)
78.
- ^xy^ + Sy\
- 8xy + 24y') dy
- 3a;2) di/ + (2'^xyV - ^x^y) dz
+
- 1/^} da; + {2xy In (^//a;) + xy} dy
d(UV)
(a)
d(tan-' 7)
x^y
(x^
6a;2/z) da;
(2//a;)
becomes
-r-
dp
a;
a"M
p cos ^,
+
xT
^P
,
av
1 a-y
- JT'
p d,p
aw
" Z
P P
2/
"^
T~
dp
P sin 0,
1 a^w
Ta 5:12"
_
"
du
XT
d<j>
the equation
.
"
d'FlSu',
^H-^P-
PARTIAL DERIVATIVES
6]
131
Find
(a)
Ans.
(a) (y
85.
86.
87.
F(x, v)
If
84.
xu^
If
If
dy/dx and
y\ 2yu
- ~ FJFy.
dy/dx
d^y/dx^
(6)
~ xy(y^ - x),
+v =
prove that
0,
x^
if
-2xyl{y^
(h)
- xv'' =
f(x,y),v^gix,y) are
4x,
+ y^-Zxy =
- xY
find
(o)
^
^
(6)
0.
- ^'''uV + 4
Gx^uv^ + 2y
^'
Ans
(a)
32/
'
g + gg
2a;M''
^'
'
+ Sy^
+y
Zx^uv^
1.
If
f(x,y,r,s)
What
independent.
89.
If F(x, y)
0,
,,,ix,y,r,s)
gg + g|
prove that
,,
show that
^x.Fg
2F.,F.Fy
p3
dx'
90.
9L
^^'
^^'
94.
^^^
Evaluate
U F=
= Zu^-uv,
F(u,v)
2x^yz,
(a) If
=.
"^
5(m,i;,w)
If
+ ^^ + ^'
fix, y,z)
and
G(u,v)
,;.^
^''^
"
if
so find
and
^'^^
^"1
y, z)
0,
Ans.
(a)
+ ixyz
at (1,-1,0).
+ .,
2/
prove that
^^,1;,^),
V,
(a) in
+x -
dy
dz
d{f,g)
d(f,g)
d{z, x)
d(x, y)
__
is
Ana. 10
a functional
^('''^'^) d{u,v,w)
S{u,v,w) d{x.y,w)
is
a functional
^ ^Provmed
_-^;hh
terms of transformations.
valid.
is
w, tmd
{a)
~,
(b)
- Zv^
show that
S{y, z)
'^
l&uv^
it.
y-rz
^gj^
^(f'9)
y and
Ans. 24u^v
evaluate
g{u,v,w),
^.
+ v\
2uv^
= 2z^-xy,
=/'
and g{x,
0,
+ FyyFl
= xVT^^+j/VT^^,
dx
define x,
+ ?'' + ^^ and
= 0^ +
connecting F, G, and H. and if so find it.
x=f(u,v,w),y =
^(=^.2/.^)
95.
if
+ Zy^-~.^G =
=^^.2'
I!l
relationship
- iyz - 32a;'z'
(1 + i^y^Y
16a;'j/
(c)
^^^
v and w.
ISj/'z
<
~,
(")
- &xz - Z2x^y^
+ Sxyzr
(1
^^'
^i'^Ti^^
which
the region
^"''"*'
iTTtr-T^'^
99.
SS-
^''^
(a)
0,
^""^P""-" *^ "^^^"'^
(''^
and
[CHAP.
PARTIAL DERIVATIVES
^32
\<)(x,y)\
\d(u,v)\
Ans.
sin^Mcosh^o'
(b)
cos^Msinh^f,
(sm^ucosh^i>
(c)
m sinh
cos
^)
^u
region
- - 2 +
o _L o - .
TO
z = 2m - 2v + w. (a) Sketch the
w, .,
+w
y -- u Zv
Given the transformation x = 2u + Sv
^
x
x
by
8,
0,
j/
bounded
0,1/
4,
of the a.?/, space
%' of the uvw space into which the region
with the ratios of the volumes
(c) Compare the result of (6)
,
101.
l'o4^^0,
=6
of
102.
is
mapped.
and
Compute
(6)
Ares.
"iR'.
^J^.
(6)
x = rsincos.,
Given the spherical coordinate transformation
surfaces (a) r
coordinate
the
rgO OSfiS^ OS0<27r. Describe
- b, and
c
^-T
- a, rt
(6) 9 -^kndTcIT^
Ares,
spheres,
(a)
{&) cones,
wllere
wnere
c,
(c)
planes
(c)
_
-
9(a!, j/, z)
Qi^r,e,4>)
_
"
^ ^"^
"
Prove that
105.
Expand
Ans.
106.
In
/(x, .)
.Vx^
l-3(x-l)-2(2/ +
10[1
Prove the
108. Generalize
first
e{y
powers of x
in
l)
1)Y
+ 1, up
"^
'
"
to
e{x
- 1)Y
^^^^^^
F{P,V,T)^0, prove
(a)
gy|^ Jyl^
= ~ wl'
^^^
P,V,T
thermodynamics, where
in
P\
^^
^^
that
^\
dTW dVi
^J^
dPlr
1.
= -1
110.
Show
111.
Show that
satisfies
f{u,v)
113. If
f{u,v), y
and y
^ Au,v,w),
g{u,v),
g(u,v), z
= Hu,v,^)
d{x,y)
W^
_
~
x(dz/dx)
nx + y-z,x^ + y') =
112. If
<e<
114. If
MISCELLANEOUS PROBLEMS
109. If
and .
"^ "'
/(x,.)
107.
"'^^^ ^
Expand
Ans.
'^'^''
^TTTli^W^y
satisfies
prove that
dv
x(Bz/dy)
- -j
and F{x,y,z)
d(u,v)
"^
and u
d(u,v)
/(r,s),
8ix,y)S(u,v)
3(^)3(r,s)
z.
0,
y.
S(x, v)
eiu.v)'
prove that
3(u,v)
"
T
^nere J
g^
y(Sz/ax)
1 dy
__
h(u,v)
y(dz/dy)
g(r,s).
w=
d(x^ djj^^
a{v,w)
a(,r,s)
h(r,s).
prove that
^M ^^
d(w,u) a{r,s)
^^^^-
PARTIAL DKRIVATIVES
^J
Prove that
115. (a)
116. If ..
ae
ce
+ bg
+ dg
d(.r, s, t)
"'
c^'ondL*;:
(6)
Illustrate
r:sp;ct to
120.
homogeneous of degree
by using the
X.]
B(u, V,
special case
2,
prove that
F{x,y)
{x,D.,
5^ +
(a)
,.
w)
a;^
establishing
are functions of
true
., .
the
and
*,
rule
for
the
prove that
'
d(r, s,
,-
t)
1,2,...,..
^+
2xw
^^
i!^
9p
x^ ln{y/x).
Tf
pr-,-
'^'
\
'
'
"'''^ ''
Problem
(.),
25, twice'
with
homogeneous of degree
p,
then
121. (a)
the result in
thus
+ dh
(6) Is
cf
^.(-,...,^......,.,)=0 where
oT^"^
af+bh
while ., . and
,.
^33
^=
if
+ iy =
is
1;
according
to
x + iy
e,
^^ ^-riy
transformed into
F(u +
iv)l
(u
{u
iv)'
+ iv)
th., under
.h
bioProve that
this
Chapter 7
Vectors
VECTORS
and
SCALARS
PQ from
Q called
another point
denote vectors by bold
initial point to
We
is
|A|
or \A\.
Fig. 7-1
VECTOR ALGEBRA
The operations
of addition, subtraction
numbers
The following
1.
Two
are,
an algebra of vectors.
vectors
and
are eqtial
if
A
A
their
Fig. 7-2
The sum or resultant of vectors A and B of Fig. 7-3(a) below is a vector C formed
by placing the initial point of B on the terminal point of A and joining the
initial point of A to the terminal point of B [see Fig. 7-3(6) below].
The sum C
is written C = A + B.
The definition here is equivalent to the parallelogram law
for vector addition as indicated in Fig. 7-3(c) below.
C =
A+B
Fig. 7-3
134
CHAP.
VECTORS
7]
135
J'tj'^^''^
ID
T^Tri+^+^
Fig. 7-4
!?! ff'T"'
added
to B gives A.
^"
5.
mA
LAWS
of
If A,
1.
2.
mA
VECTOR ALGEBRA
B
and
A + B=:B + A
A + (B + C) = (A + B) + C
4.
5.
m(A + B) =
3.
I=b
wA + mB
then
Law for
Law for
Distributive Law
Distributive Law
Associative
Addition
Associative
Multiplication
'"
^^^^ "^y multiplication of a vector
^''V^tJ
defined.
On Pages ^^tT
136 and 137 we define products of vectors.
is
UNIT VECTORS
thPv,
then
^^^1
AM
We
/O
Fig. 7-5
A>0
= Aa!
[CHAP.
VECTORS
136
(AijA^.A,)
Fig. 7-7
Fig. 7-6
COMPONENTS
of a
VECTOR
Any
of
A =
The magnitude
of
Aii
is
A23
+ Ask
{!)
lAl
V^fTAlTAf
{2)
or
|r|
\/x''
+ j/^ + z^.
SCALAR PRODUCT
dot B)
and B, denoted by A'B (read
scalar product of two vectors
angle
the
of
cosine
the
and
B
and
defined as the product of the magnitudes of
The dot or
is
write
is
A =
DOT
we can
In symbols,
between them.
A-B =
Note that
A B
is
3.
4.
i-i
j-j
(^)
valid:
A'B = BA (B + C) = A B + A C
m(A B) = (mA) B = A
2.
O^^Stt
ABcos6,
= k-k =
1,
Law
(mB) = (A B)m,
i-j = j-k ^ k-i =
where
is
a scalar.
CHAP.
VECTORS
7]
If
5.
A=
Aii
A-B =
A- A
If
6.
CROSS
or
A B =
and
and
B2J
+ Bgk,
then
=
=
B-B
Bii
I37
and
are perpendicular.
VECTOR PRODUCT
where u
A X B = AS sin
is
to B, then
6*
and we
AxB
1.
6>
define
AXB =
A X B.
tt
If
(5)
A=B
3.
4.
if
is
parallel
valid:
-BxA
(Commutative
Law
2.
or
0.
is
a scalar.
i,
5.
AXB
Ai
A2
k
A3
Bi
B2
B3
6.
|A X B]
7.
If
are parallel.
TRIPLE PRODUCTS
Dot and cross multiplication of three vectors A, B and C may produce meaningful
products of the form (A-B)C, A- (B x C) and A x (B x C). The following laws are valid:
1.
(A -6)0
2.
A(B-C)
in general
A
3.
4.
AX
(B X C)
Ai
A2
Bi
Ci
B2
As
Bs
C2
Ca
^ (A X B) X C
(Associative Law
(BXC) = (A-C)B - (A-B)C
(A X B) X C = (A- C)B - (B- C)A
X (B X C)
(S)
The product A (B x C) is sometimes called the scalar triple product or box product
and may be denoted by [ABC]. The product A x (B x C) is called the vector triple
product.
VECTORS
138
[CHAP. 7
AXIOMATIC APPROACH
to
Definition.
With
Thus
VECTOR ANALYSIS
seen that a vector r = xi + y] + zk is determined
some coordinate system are known. In adoptthus quite natural for us to make the following
it
is
relative to
it is
3 dimensional vector
we can
is
an ordered
2.
3.
A-B
4.
and subtraction,
etc.
6.
7.
Length or magnitude of
A =
= VAvA =
|A|
(1,0.0),
vectors, such as
etc.
By
(0,1,0),
+ A| + A|
A + B = B + A,
\/^f
A+
(B
+ C) =
k =
(0,0,1)
(7)
that
A In like
(0,0,0)
5.
numbers
if
1.
manner we can
define
AxB
All
A2J
+ A3k
(S)
After this axiomatic approach has been developed we can interpret the results geoFor example, we can show that A B = AB cos 6, |A X B] =
metrically or physically.
sin 0, etc.
AB
triplet of real
In the above we have considered three dimensional vectors. It is easy to extend the
idea of a vector to higher dimensions. For example, a four dimensional vector is defined
as an ordered quadruple (Ai, A2, A3, A4).
VECTOR FUNCTIONS
If corresponding to
is
Ai{u)i
called
+ A3{u)k.
The function concept is easily extended. Thus if to each point (x,y,z) there corresponds a vector A, then A is a function of {x,y,z), indicated by A{x,y,z) Ai{x,y,z)\ +
A2 {x, y, 2:) j
We
+ A3
(a;,
y, z)k.
sometimes say that a vector function A.{x,y,z) defines a vector field since it
associates a vector with each point of a region. Similarly 4>{x,y,z) defines a scalar field
since it associates a scalar with each point of a region.
CHAP.
VECTORS
7]
139
of VECTOR FUNCTIONS
Limits, continuity and derivatives of vector functions
follow rules similar to those
for scalar functions already considered.
The following statements show the analogy
which
exists.
1.
ever
The
2.
\u-ua\
<
This
S.
derivative of A(tt)
is
is
lim A{u)
positive
e
when-
A(mo).
defined as
dA
^
du
+ A^<)~A(^)
A(^<
lijn
Au-0
In case
A{u)
{u)i
+ Aa {u)j + As {u)k;
then
dAs,
du
du^
Higher derivatives such as d^A/du^,
3. If
is
A(a;, y^z)
du^
etc.,
'^
'W^
then
the differential of A.
4.
are:
|^(M)
^(A-B)
(c)
|(AXB) =
(a)
GEOMETRIC INTERPRETATION
'^
of a
du
du
dy
,
^ aB
'
dy
,
aA
VECTOR DERIVATIVE
of a
coordinate system and the point {x,y,z), then
specification of the vector function r{u) defines
X, y and z as functions of u. As u changes, the
terminal point of r describes a space curve
(see
Fig.
x{u),
7-8)
hailing
y{u), z
parametric equations
If the parameter u
the arc length s measured from some fixed
point on the curve, then
z{u).
At
is
ds
(9)
dr
is
^
(io)
di =
the velocity with which the terminal point
Fig. 7-8
r(M
+ Am)
r(M)
[CHAP.
VECTORS
140
We
from which we
have
_
~
_
~
dr
di
dtds^
ds dt
magnitude of v
ill)
dt
dt^
is
Similarly,
ds/dt.
'
These con-
V
Then
if
.^(ic,
y. z)
which
dition
1.
is
many
z)
grad ^
(13)
have continuous first partial derivatives in a region (a concases stronger than necessary), we can define the following.
and A{x, y,
in
by
defined
(del)
'^^^dy +
V*^
defined by
is
<j>
dz
dy
dx
{U)
^dz
of a
interesting interpretation is that if ^{x, y,z) = c is the equation
Problem
36).
(see
surface, then V^ is a normal to this surface
An
2.
VA
divA =
[^Tx
is
Ml
dx
3.
Curl.
The
curl
curl of
A =
is
defined
A =
'^
defined by
-^
(All
^Jz
'S^J
dAi
dAa
dy
dz
A2J
+ Ask)
{15)
by
i^ + ^4 +
i
dy
{16)
^^'^^
A
A
dx dy
Ai
*^^)"^^^^ + ^^^ +
1dz
Ai As
d
dz
dx
_5_
dz
dx
dy
+ k
A, A2
A, As
Ms
dA
dy
dz
dAi
i+{^-^n
dz
Ai A2
Ma
Ml
5a;
dy
Note that in the expansion of the determinant, the operators d/dx,d/dy,d/dz must
precede Ai, A2, A3.
CHAP.
VECTORS
7]
141
FORMULAS INVOLVING V
If the partial derivatives of A, B,
1.
2.
3.
4.
5.
6.
7.
8.
9.
11.
12.
and
+ F) = vf/ + VF
or
grad(f/+F) = grad 2* + grad F
V(A + B) =
+
or
div(A + B) = div A + div B
VX(A + B) = VXA + VXB
or
curl(A + B) = curl A + curl B
V{UA) = (VC/)-A + f/{vA)
V X (C/A) = (Vf/)XA + Z7(VXA)
V(AXB) = B-(VXA) - A-(vxB)
VX(AXB) = (B-V)A - B(VA) - (A-V)B + A(VB)
V(A-B) ^ (B-V)A + (A-V)B + BX (VXA) + Ax (VXB)
V(C/
VA
V.(vf7)
V
V
VB
v^t/
V2
and
10.
^ + ^,
32
52
is
52
.
a^ ~^ 5^2
5a;2
is
X (vt/)
(V X A)
f{ui,U2,U3),
g{ui,U2,U3),
h{Ui,Ui,us)
(17)
xi
y]
zk
f{uuU2,uz)i
g{ui,U2,Us)j
h{Mi,U2,U3)k
(18)
{ui.m.Ua) the
point.
r describes a curve
From
dr
{18),
we have
- dui
aUi
dU2
dU2
Fig. 7-9
(19)
dUa
dt/du^ is tangent to the u^ coordinate curve
at P.
direction, we can write dr/du,^h^^, where
=
The vector
this
write
tively.
dT/dU2
Then
and dT/du^^hzez,
can be written
h2e2
(19)
dv
The
hiduiei
where
h2
If ei is a unit vector at
h^
\dTlduA.
\3r/dU2\
and
h,
Similarly
\dr/du,\
\
hzduzez
hsduaca
we can
respecy
(20)
VECTORS
142
If
[CHAP.
ei, 62, es
called orthogonal.
and corresponds
dx-dx
h\du\
h^dul
h^dul
{21)
above parallelepiped.
ds^
dV
{hi
dui
ei)
is
given by
{22)
dV
where
dx
dx
dx
dUi
dUi
dUa
d{x,y,z)/d{ui,U2,Ua)
d{x,y,z)
{23)
is
It is clear that when the Jacobian vanishes there is no parallelepiped and explains
geometrically the significance of the vanishing of a Jacobian as treated in Chapter 6.
in
ORTHOGONAL
CURVILINEAR COORDINATES
If * is a scalar function and
curvilinear coordinates Ui,U2,Ua,
1.
V*
V* A
div
A =
curl
ei
Jhi dUi
A =
Aiei
grad $
A =
we have
{hihaAi)
dUi
'
dU2
hiCi
^262
haea
dU2
dUa
hih2ha
dUi
Laplacian of *
V'*
hih2ha
{hahiAi)
{hihiAa)
h2A2 haAa
hiAi
4.
ha dUa
A =
es
'
hi dUz
T
hihiha
62
J-
dUi
a
a$
/ /12/13
\
hi
/hahi a*_
dU2 [
dUi
hi
dU2
f hih2 a*
dUa\^
ha
dUa
{ui, 2, Ua)
if
hi
we replace
= h2 = ha = 1.
Cylindrical Coordinates
P(x,v,z)
Transformation equations:
X
where
p cos
0,
Scale factors:
^ <
27r,
hi
1,
Jacobian:
d{x, y, z)
- f
-tt
d{p,
<j>,
<^,
z)
Element of volume:
psm<j>,
< .
h2 = p, /13 = 1
ds^ = dp^ + p^ d<j>^+
=0
<
2;
'^
dV =
pdpd<f>dz
Laplacian:
^ ^ -
~pdp[p'd^J
dz^
7a^
a^iaC7^^_^
'W
dp^
P dp
"^
p2
a^2
9^2
CHAP.
VECTORS
7]
143
2.
Spherical Coordinates
(r, d, <f).
Transformation equations:
<^,
ir,S,<!>)
,r,
dr^
'^^
JacoUan:
Element of volume:
r^sin^ed^^
r^^inS
Pig. 7-11
dV =
^^U =
Laplacian:
r^de^
dr dd
r^ sin 6
r^dr\
r^
dr J
d<i>
J-_ ^
r^smOdd
Solved Problems
VECTOR ALGEBRA
1.
Show that
and
Then
addition of vectors
is
commutative,
OP + PQ = OQ
OK + RQ = OQ
A+B = B+
i.e.
or
or
A + B = B + A.
A + B = C,
B +A = C
A.
Fig. 7-12
2.
we have
(A
B)
OP + PR = OR =
OQ + QR = OR =
A+
(B
C)
vec JsTiiTaii!;'
(A + B)
'"""^
and
D,
D,
i.e.
-I-
(B
+ C) = (A + B)
PQ + QR = PR
i.e.
A + (B + C) = D
i.e.
(A + B) + C = D
(B
-f-
C.
See
+ C)
C.
'"^^ '-'^^'-
"' ^'^'''^^
'^^
-y -->-
of
VECTORS
144
[CHAP.
An
automobile travels 3 miles due north, then 5 miles northeast as shown in Fig. 7-14.
Represent these displacements graphically and determine the resultant displacement
(a)
graphically,
Vector
(&) analytically.
OP
or
PQ
or
north.
Vector
east.
(6)
C'
and
A'
C=
By
B'
- 2AB
Thus vector
Prove that
5.
If
same
OQ
5^
sin
sin
A
lOQP
lOPQ =
+ yb
+ yih =
X2a
+ y2h,
34
C
sin Z
55.21
3(0.707)
.
Then
OPQ
ookk
0.2855
implies
v^^here a
xa
10QP =
a
and
(45
1635')
+ yh =
ABCD
Wz =
0,
2/1
Problem
implies
0.
and b are
1635'
xa = -j/b
non-collinear, then
3/2 =
or
{xi
or xi
X2)a +
xz,
2/1
{yi
a;i
y2)h =
2/2.
49).
intersecting at
i.e.
mile
I5V2
-I-
X2a
Let
Fig. 7-14
Then xa
Hence by Problem
6.
3"
Xia+2/ib
Xia
lOPQ =
if
Suppose x=0.
to the
cos
7.43 (approximately).
A
,
sm lOQP =
4.
Unit
x(b
a).
= X2 and
1/1
= 1/2.
CHAP.
VECTORS
7]
7.
From
and CB.
sides of a triangle
AC + CB = AB or b + a = c.
be the line joining the midpoints of sides
parallel to the
is
Fig. 7-16,
DE = d
Let
two
145
Then
= DC + CE = lb +
Thus d
la
l(b
parallel to c
is
AC
length.
its
Fisr.
8.
AA + Aij+Ask
By
where
A of the vector A =
A = A/Af+Af+Af. See Fig.
is
_
OP
{OQY + (QPy
A,]i
= (ORf + (RQ)\
(OPy = (OBY + {RQy + (QPy
(OQ)'
AI
AI
A\,
or
A = VAH^4|TAf.
i.e.
9.
tude.
P
Q
position vector of
position vector of
+ PQ =
PQ = r2 ri
=
=
xii
iC2i
+ yij + zik.
+ 2/2J + z^k.
Q(.Jt-s,
ri
- xi)2 +
V(a;2
or
is ra
or
=
Note that
is ri
r2
Magnitude of
DOT
i'{xi,yi,z,)
The
The
The
<{
A-
iOPr
Then
(2/2
between points
Fig. 7-18
SCALAR PRODUCT
A on B
is
is equal to A-b,
a unit vector in the direction of B.
Through the
perpendicular to
Fig. 7-19; then
Projection of
11.
P(A...A,,A,)
Similarly,
10.
7-16
Prove
initial
at
G and
B =
on
GH = EF = A
A-(B-fC) = A-B
-h
-h
(B
C) on
A =
= A-b
cos 9
Fig. 7-19
A-C.
pass planes
projection of B on
+ projection of
then
A
C on
Multiplying by A,
(B-f-C)
and
(B
C)
Aa
-
= B-Aa
A = B A
-f-
-I-
C-Aa
C
A-(B + C) = A-B4-A-C
and the distributive law
is
valid.
G A
Fig. 7-20
Vt, 2)
[CHAP.
VECTORS
146
D.
0-0-0 =
13.
=
=
=
(a)
i-i
(b)
i-k
(c)
k-j
id)
i(2i-3j +
(e)
(2i-j)-(3i + k)
14. If
A =
|i|
= (1)(1)(1) = 1
= (1)(1)(0) =
= (1)(1)(0) =
cos 90
= 2j-i - 3j-j + j'k = 0-3 +
= 2i-(3i + k) - i-(3i + k) = 6i-i +
cosO
|i|
|i||k|cos90
HlJl
Aii
k)
Bii
= -3
-3ji-jk =
2i-k
+ B23 + B3k,
prove that
A-B =
A1B1 + A2B2 +
A3B3.
A-B =
=
=
since
15. If
i'i
A =
\/A\
+ A| + A
+ {A3){A,) = A\ + A\ + Al
B = A.
Then A = VAvA = VAf+Af+Af is the magnitude of A.
=
by Problem
Prove
(A2)(A2)
14, taking
The CROSS or
16.
{A,)(A,)
Sometimes
A A
is
written
A^
VECTOR PRODUCT
AXB
-BxA.
IS
(6)
(a)
Fig. 7-21
AXB = C
system
has magnitude
AB
sin e
B X A = D has magnitude BA sin e and direction such that B, A and D form a right-handed
system [Fig. 7-21(6) above].
Then D has the same magnitude as C but is opposite in direction, i.e. C = D or A X B = B X A.
The commutative law for
CHAP.
17.
VECTORS
7]
147
AxB AxC
Ax(B
Prove that
+ C) =^
+
for the case where
is perpendicular to
B and
also to C.
Since A is perpendicular to B, A X B is a
vector perpendicular to the plane of A and B
and having magnitude AB sin 90 = AB or
magnitude of AB. This is equivalent to multiplying vector B by A and rotating the resultant
vector through 90 to the position shown in
Fig. 7-22.
Similarly, A X C is the vector obtained by
multiplying C by A and rotating the resultant
vector through 90 to the position shown.
AX(B
18.
AXB-FAXC.
AXB
+ AxC
B and C are
non-coplanar.
Fig. 7-22
B^
= B
is
sin
the
e.
Also, since
Now
B+ C =
B,
B,,
AX
Fig. 7-23
and
so
by Problem
it
follows that
17,
+ Cx) = A X Bx + A X Cx
AX(B + C) = AXB + AXC
Then
(Bx
and the distributive law holds. Multiplying by -1, using Problem 16, this becomes (B -1- C) X A
A.
Note that the order of factors in cross products is important. ~'
The usual laws of
BXA + CX
"
if
proper order
is
maintained.
19. If
A=
Aii-I-Aaj
AXB
=
=
=
"
+ Aak and B =
Bii
+ Bzj+Bah, provethat
AXB
AAX
Si B2 Bs
Ai A2 As
Hi B2 B3
20. If
A =
3i
- j + 2k and B =
i
2i
3-12
AXB
- k,
Sj
-1
5i
7j
Ax B.
find
-1
3-1
21.
[CHAP.
VECTORS
148
-1
+ k
-1
+ Ilk
=
=
=
of parallelogram
|B|
|A|sin9|B|
lAXEJ.
and
B = llAXBl.
22.
Fig. 7-24
Q(4,2,-l),
Find the area of the triangle with vertices at P(2,3,5),
PQ = (4 - 2)i + (2 - 3)j + (-1 - 5)k = 2i - j - 6k
PR = (3 - 2)i + (6 - 3)i + (4 - 5)k = i + 3j - k
= 1 |(2i- j - 6k) X (i + 3j -k)
= 1 PQ X PR
Area of triangle
/2(3,6,4).
k
-6
-1
13-1
iV(W+T^WTW
19i
- 4j + 7k
iV^
TRIPLE PRODUCTS
23.
(B X C) is in absolute value
that
of a parallelepiped
volume
equal to the
Show
with sides A,
B and
^izxi^
C.
parallelogram
Fig. 7-25
/.
Volume
of parallelepiped
(A'n)(|BXCl)
= A
If A,
24. If
B and C
{|B
C| n}
= A
/)
(B X C)
A n <
(B X C)
Ai
A2
Bi
Bi
Ci
C\
i
(B
C)
= A
Bi B%
C\
Ci
= A
(B
A3
Bs
Ca
k
Cl
{B,C^-B.C,)}
+ {B.C^- B,Ci)k]
Ai A2 As
= A,{B^C,-B,C^)+A^{B^C,-B,C,) + A.{B,C.~B.Cd
C)
show that
B:
Bi
B. B,
C,
C2
Cs
CHAP.
25.
VECTORS
7]
149
3i
B =
j,
C =
2k,
5j
4k.
3-10
volume of parallelepiped
24,
(B
12
15
C)
i
-201
26.
Prove that
(B X C)
= (A x
B)
C,
i.e.
24:
(B
C)
At Aa
By Problem
Ci
(AXB)-C =
Bi B2 B:
Ci
20.
C2
C-(AXB)
Cs
C2 03
Ai A2 A3
Bi Bz Bs
27.
Let
ri
and
rs
and
Find an equation for the plane
passing through Pi, P2 and Ps. See Fig. 7-26.
P3(a;3, ya, zs).
=:
xi
in the
plane.
+ yi + zk
vectors P1P2 =
r - ri which all lie in the plane.
PiP-PiP2 X PiPa
or
(r
- ri)
(r2
- r,) x
(ra
/'
Then
=
- ri)
Fig. 7-26
or,
28.
(g,
_ g^)^
position vectors of
ri
''^^
.oK
equation
IS
Si
+ j - 2k,
rs
-i
2j
4k,
ra
" ^ ~/
^'^' 7 '' ~
^'^' " "' ~ "''
T'
(r - n)
(r^ - n) X (ra - n) = 0,
{(X - 3)i +{y- l)j + (z + 2)k}
{-4i +
""''
""''
2i
- j + k,
xi
-2), P2(-l,2,4),
1,
respectively
yj
zk
i.e.,
{{x
Another method:
Pi(3,
By Problem
6j
-h
6k}
9k}
or
X {-i -
2j -h
3k}
=
5x
2y
Sz
11
5x
2y
3z
11
VECTORS
150
A = +
29. If
B = 2i-3j+k, C = 4j-3k,
j,
AXB
(a)
BXC
(b)
[CHAP. 7
i - i - 5k.
-3
(A X B) X C =
Then
--1
k
-5
-3
-3
-3
5i
6j
AX
Then
8k.
(B
(A X B) X
AX
(B
23i
+ 3j + 4k.
110
C))
(&)Ax(BxC).
(A x B) X C,
(a)
find
= 8i-8j+k.
C).
DERIVATIVES
30. If
t
(a)
(i3
+ 2i)i-3e-2j + 2sin5k,
^(t'
At
(6)
From
dt'
2)i
tion
|dVd'|
(3J^
+ 2)i +
6e-j
10 cos5
/lio
6e-"j
at
2\/35
10 cos 5f k}
0.
6ti
12e-"3
50 sin 5t
-12j.
From
(c),
at
df
10k
(d)
If
31.
{(3i^
dt
= 0, dVdi'
6]
(b)
significance.
VW+~W+JW
dt\dtj
At
2i
fdr
_
~
d^r
(c)
dridt
\dtldt\
(a),
^,
(a)
= 0,
dh
dr
find
12
at
0.
represents time, these represent respectively the velocity, magnitude of the velocity, accelerat =
of a particle moving along the space curve x t^ + 2t,
-3e-^', z
2 sin 5t.
Prove that
-p-
(A B)
dB
dA
\-
-^
-j^
where
B,
and
are
diff erentiable
func-
tions of u.
Method
1:
;v-(A-B)
du
+ AA)
(A
lim
(B
lim
Am
A-o
hm
A-o
Method
2:
A =
Let
d_
du
Aii
x^yz and
(x^yz){Zx'^yi
(Sx^y^zi
dz
'
Ask,
B =
Bit
B^j
_|_
dM
dA^
dM
Then
Bsk.
du
_
~
0A =
^ dB
(A-B)
'
\^
A2J
(a
""
32. If ^{x,y,z)
AA
AA ^
AB
^^
A
+ -r-'^
+ 7
Am
Am
Am
/.
,.
+ AB) - A-B
Am
A->o
;^
^+ ^+A
A.
dM
dB ^ dA
du
dM
A =
yz^i
dM
+ yz^ -
Kzk)
^ +
du J
(Bi
\ dM
'
dM
^2^
dM
Bs"^
2,x^yi
xVz'j
3a;V^zi
x'yz^'k)
xz^i,
find ^r^(</'^)
dydz
x'y'^z^i
3x*yH
x^yz^k
Sa^j/^gSj
2x^yzk
CHAP.
VECTORS
7]
dy dz
If
33. If
(3a;Vi
dy
= l, y~-2, z- -1,
A =
x2 sini/i
Method 1:
3A _
;r
(0A)
z^ cosi/j
sm yi .
2a;
,.
dA
Method
aA
=
=
=
aA
y' k,
gy
(2x
(2x
xy^k,
a;''
cos
find
dA.
2^ sin
3/ i
2/
=
=
sin
dix''
(2a;
sin
+
+
2/)i
y dx
x^ cos
^^
2xyz''i
^^^
+ xVi +
(i+^^
aa;
(^^)
22 cos yj
j/ j
y dz
y dz
z^
+
-
2xyk) dy
sin
d2/)j
j/
(2z cos
(^/^
j)
da;
dz
2a;j/
dy)k
(2z cos
find
z^
sin
v^,
(a)
dj/)j
j/
(6)
div (0A)
d/dy
B/dz
xz
y^
2x^y
curl (5SA)
9x
A,
da;
(c)
|-
2xy dy)k
V X A,
(a;^2/2)j
(d)
2x^yk)
^(2^'2/)
(a;
(^A)
^x^yz*
(x''yz*i
V X (0A)
VX
(x'yzH
d/dx
d/dy
d/dz
x^yz*
-xVz^
2a;V2'
(4a;Vz'
+ ^(-a;yz) +
Zx'y'z' + 6x''yV
^{x'yz')
^ {x^yz^Yi
"
+|k).(a;zi-,^i +
d/dx
2a;^i
Mj+Mk
By
dz
(j/"
22/
+ h^ + ^k)x(a;zi-2/^j + 2x^2/k)
(^*
Zx^yz^k
^(-2/')
(4<^^^^)
(e)
dA
2a;2/k,
curl (^A).
(e)
(d)
d(a:2/^)k
dy)i
2/
Mi
dx
(">
2k,
j/ j
sin
z''
(2z cos
a;^
(6)
2x'zk
2:
dA
(^A),
12j
Bx^yz'j
2/
2/ i
Gx'yi
-12i
3A
flA
+ ^-dj/
+ -T-dz
dy
dz
sin i - y^k) dx + {x^ cos
sin
da; +
cos y dy)i +
,
-^-dx
ox
2x'yzk)
becomes
this
9a;
Sx^y^z^
151
3a;Vz')i
x^y'z^i
^(2.^^))i
^ i-y') + (&
2x*y'z^)
^(2a;2/V)
xVz^J
+ (4a;V -
2x''y'z'^)
8a;Vz')J
(2a;2/'2'
a;V)k
^(<^'')y
div
(a;2/z)k
VECTORS
152
35.
Prove
(0A)
(^A)
'
(V^)
{</>Aii
A +
[CHAP.
^(V A)
+ ^Ash)
'
0A2J
dz
dy
aa;
dy
dx
dz
(V0)-A + 0(V-A)
36.
V^
Prove that
is
<j,{x,y,z)
c,
where
c is a
constant.
Let
Then dx
i.e.
37.
V0
dr
xi
V^
so that
is
Find a unit normal to the surface 2x^ + ^yz By Problem 36, a vector normal to the surface is
V(2x^
2x^y
(6)
=
=
VV
to the surface at
is
dy
and
V^
(a)
=
=
'^ct>
at (3,-1,2)
= 12i + 8j - 24k
12i + 8j - 24k
_ 31 + 2j - 6k
(-24)"
(8)='
+
V(12)^ +
31 + 2j - 6k
10)k
V^^-
(&)
(ixy
dz
'
Laplacian of
Ay
(4^/
is
dx
4zj
find
xz^,
4x1
to the surface at
38. If
+ 4:yz-hz^)
= -10
5z^
z^)'\ +
=
2*")
3x2^k
6xz
Another method:
_
-
5V
dx-
Ay
39.
Prove
dy-
^{2x-y-xz')
dz-
A =
0.
(V X A)
\.
>
d_
dx
assuming that
immaterial.
+ ^{2x-y-xz') + ^{2x-y-xz'>)
Gxz
div curl
div curl
/ a A _ 5A^\
dz J
\dy
d-As
d-Ai
dxdy
dx dz
a/aa;
d/dy
d/dz
Ai
A2
A3
fdA,
^A3^.
dy dz
dy dx
aAj
dxj'
aAsN
d /sAi
\
A/'
Ml _ 3A3
dy\dz
dx J
3M3
aUa
d-Ai
(,a.
dz dx
_ 5A
3a;
9_fdA2
dz\ 3x
0^]
31/
dA
dy
d^At
dz dy
is
CHAP.
VECTORS
7]
JACOBIANS
40.
Find ds^
(a)
Method
CURVILINEAR COORDINATES
and
in
(a)
and
cylindrical
(6)
a;
Then
= -p
=
ds^
Method
= h^^ 1,
The
2:
sin
c?0
dx^
h,
position vector
dx
dp
(cos
dr-dr
a;
Then
= -r
cos
(p
<b
cos
p cos
p sin
<p
dp,
dz
dz
dp)^
<t>d<i>
sin
sin
rfp)^
(dg)^
^,^^^^,
Then
zk.
Sz
sin0j)rfp
=
=
r sin
IS
fl0
(cos0i
(6)
P cos
(- p sin ^d,p
dz""
=z
d,f +
p sin 0,
p^(d^Y
dp
ds^
2/
dy
dp,
(dpY
=
=
Thus
dy''
P cos 0,
cos
^
h,
1:
dx
and
153
p sin
(cos
(dp)'
ff
d0)i
<f,
cos 0,
sin sin0 d0
j/
r sin # sin 0,
r cosff cos0 de
(dz)2
r cos 9
sin 9 cos
dr
dy
and
41.
dV
(a)
dV
in (a) cylindrical
and
(6)
in orthogonal curvilinear
coordinates
In cylindrical coordinates,
h.h.h.du,du.du,
Ui= p,
p,
dV =
U2i
d,
-P,
u!,
us
(l)(p)(l)
is
\^ip.l^\du,du.du.
= ^,
z ni
h,-
dp d0 dx
\
1,
h,~
/i2-p,
,
h.3
-1
r
[see
r.
u,
Problem
./ .,
40(o)].
Then
pdpd^dz
/'
/(' =
J=
(r
sin
r^ sin
(9
(o)
Volume element
in cylindrical coordinates.
(6)
Fig. 7-27
Volume element
rf0)
(r rf6') (i/r)
sin6'
rf<^
in spherical coordinates.
[CHAP.
VECTORS
j^54
(6)
In spherical coordinates, Wi
= r,
dV
= e,
m.
Us
= <t>,h, =
(l)(r)(r sin 0)
dr de
1,
d<p
h^^r,hz =
r sin e
dr de
r" sin e
[see
Problem
40(6)].
Then
d<p
42.
Page
(a)
= p,
M2
142.
Then
grad*
where
ei, ez, Cs
= 0,
W3
= 2,
(c)
A =
V^*
=
"
j ^^^e.
fe
= P,
^^
Aiei
IK
<'-.'
=1
[see
(b)
div A,
Problem
19*
e.
Cs
in the results 1, 2
40(a)]
9*
-^e,
V^*-
(c)
+
p,
S*
and 4 on
5*
+ -ea
,
e.
0, 2 respectively.
^]
+ ^262 + Aaea.
riA^Ki)^*^
(l)(p)(l)
*,
/is
15*
grad
=
where
= 1,
a*
V*
fei
(a)
dpV
Lap\
dpj
dp
(1)
"^
p'a0^
50\
J
"^
a /(i)(p)
i^Al)a)a*\
90/
(p)
92\
9*\"|
9z/J
(1)
92"
IMISCELLANEOUS PROBLEMS
where
43. Prove that grad /(r) = ^r,
yx^+^M^
and
/'(r)
d//dr
is
assumed
to exist.
grad/(r)
V /(r)
/'(r)^i
Another method:
f'(r)^j
/'(r)^k
^(xi + yj + zk)
we have
a*
hi dui
a*
h-t
du%
9*
/,N
hi 3u3
* = f(r), a
'^
44.
(ft)
equation
V^^ =
0.
<^
'
/(r).
(b)
dr
we
letting
have, on
Prove that
Then
<i,
= Vr
is
a solution of Laplace's
*^HAP.
(a)
VECTORS
7]
By Problem
43,
V0
By Problem
35,
Laplacian of ^
Another method:
assuming that
(6)
U = f(r),
From
Sry
"^
(V96)
is
we have
V-J^^rl
we have
ao/ +
r^sinffSeV'"
we
r= sin^ e
~d^
find
we have
(a),
v{^)
\rj
45.
^^r
r
f(r)
V/(r)
In spherical coordinates,
r^arV
ii
I55
^(l\
_
~
dfl\
dr^\rj
rdr\rj
2
r"
_
-
dv
where T and
v^
rf^r-
fa^yV
r/d.^Y
,
ds^
The
dv
dt
Since
T has
_
- dt^ +
dt^"^^
unit magnitude,
^^^*
^'lF
,''
1.
_
-
dvdT ds
di^ + ^'Tsdi =
Then
2X.^ =.
-^T-^
u,
0,
/d^z^''~"'
+1^
vt.
dT
we have T T =
+ -- =
rfj-r^-x
''S
tfTrTll'V^
^^iT.u^^^
dT/ds, and called
01
the principal
by v
di^
A/T^
^\d^)
ds^J
"^
dv
d^T +
dT
T-^
or
dF
given by
dT
^^f
P^'-P'^di"l^r to T. Denoting by
to the space curve, we have
normal
is
s,
(2)
[CHAP.
VECTORS
156
where
dT/ds
k is the
Now
magnitude of dT/ds.
f/
=
Defining p
= 1//C,
(2)
becomes dT/ds
N/p.
dr/ds
\2
,,
{(S)
ISI
T =
since
= dVds^ Hence
/j2
(S)
- (i?)
(i)
we
Page
(9),
\0
/^2 \^
Thus from
^T
dt
equation
[see
we have
139],
*"
have, as required,
^N
p
T and N
The
acceleration.
components of the acceleration, the latter being sometimes called the centripetal
of the space curve.
curvature
and
curvature
radius
the
of
respectively
are
and
k
quantities p
Supplementary Problems
VECTOR ALGEBRA
46.
47
A man
48.
If
50
and B,
4A +
3(B
- A) = A +
3B.
is
49.
prove that
are any two non-zero vectors which do not have the same direction,
and
B.
by
determined
plane
the
in
lying
a vector
and
mA + nB
A,
and
plane)
ABCD be any quadrilateral and points P,Q,R and S the midpoints of successive sides. Prove
PQRS is a parallelogram and (6) that the perimeter of PQRS is equal to the sum of the
that
(a)
lengths of the diagonals of ABCD.
Let
51.
52.
C = 3i-2j +
in
4k.
SCALAR PRODUCT
(A-B)|
|(A + B)
DOT
53.
Evaluate
54.
A =
2i-i + k,
B =
+ ] + 2k,
Ans.
The
or
is
if
A =
21
- 3j +
[Hint:
5k
and
B =
Take the
3i
+ j-2k.
sides as A, B,
Ans. 24
C where C =
A - B.
Then
C-C = (A-B)'(A-B).]
55.
Find a so that
56.
If
A =
2i
21
- 3] + 5k
and
31
+ i + k, B = i-2j + 2k
direction of B.
Ans. 17/3
aj
- 2k
and
are perpendicular.
C = 3i-4j +
2k,
find the
Ans. a
-4/3
projection of
A+C
in
the
^^^^-
"
7]
VECTORS
-j^g,^
^''''
tro^Tl'^TelT^!t)1''-'y^~y^^^^^
()
the
58.
59.
The CROSS or
60.
If
A =
"'
In'
+ BA)/(A + B)
of the
median drawn
and B.
VECTOR PRODUCT
2i-j + k and B =
tS+W^'
AXB = AX
2j
- 3k,
find
+ B) X
|(2A
(A-2B)|.
62.
If
63.
U. Find^the
len^h
C,
B=C
does
shortest distance
A =
31
Ans. 25^3
-2i + 4k and B =
+ i - 2k.
necessarily?
Ans.
by
^^3
(1,1,0),
(3,-1,1),
TRIPLE PRODUCTS
66.
Prove that
(a)
(6)
AX
= B (C X A) = C (A X B)
(BXC) = B(A-C) - C(A-B).
X
(B
C)
''
68.
69.
Prove that
(A X B)
(C
X D) +
(B
C)
^'''"^^
(A X D)
(C X A)
Ans. |3
(B
X D)
0.
DERIVATIVES
71.
72.
73.
Prove that
AX^
du
:^(AXB) =
du'
^vn
+
^ du^^
a cos .
b sin
yz]
()
.t,
u
^here
a;
= f.
,y
2/
and
- f^
^
/s
t'
at +i,
the
4-
pomt. where
,
+ xyzk,
(a
find
b),
1^
(6)
_|
.^r
0.
Ans. leVE
= 1.
is
a constant
[CHAP.
VECTORS
158
If
77.
If
4,
B have
U, V, A,
VU+VV,
(a)V(U + V) =
V -(A + B)
(6)
+ yz + zx and A =
xy
at the point
-1,
(3,
(a) 25,
where
78.
Show
79.
Prove:
80.
Prove that
81.
82.
If
83.
(a)
that VX(r''r)
VX
(a)
A =
(C7A)
yzj
V X
Prove that
Problem 82.
(Vt7)
U = 0,
grad
curl
Sxz^i
(x
+ y'zj + z'xk,
x'yi
Ans.
2).
V-A + V-B,
:=
(6)
xi
A +
find
A),
(a)
and r
(6)
V X
V0,
(A
+ B) = V X A + V X B.
V A
(6)
and
(c)
(V<p)
- 30] + 47k
56i
(c)
+ z'k
yj
t/(V
2/
(c)
|r|.
(A X B)
= B
(V X A)
- A
(V X
B).
+ 2z)k,
- 2xz + 2y^z* =
= - V^A + V(V
(V X A)
Ans. - 6a;i
curl curl A.
find
A).
at the point
10
(b)
(2, 1,
-1).
- l)k
(62
(a) if
is
given as in
85.
Express
(a)
l^ir^Ai)
1
^r_
dr
8U2
dU3
V^*
in spherical coordinates.
3*
1
;_
o<p
sm aa^^^
(smeA2)
/ ,a*\
(0)
dr
div A,
(6)
,19*
+ -^-^2
+
r ad
(b)
S'i'\
^-^
AiCi
+ AiCi + Ase^
5^
f
+ ^s-ihT^^l^sms
^^y + ^s^^2gQ^2
,
v^^\J"a^J
X{v} - r^),
factors,
Ans.
Write
Ans.
(c)
(6)
ds^
(c)
u^
(a)
=
v",
(u"
{u"
V"* and
(a)
V^*
(6)
div
div
1
dz\
hi
- y/vFTV,
h2
h^
/s^*
^7+^(^5^ +
(a)
z.
a^<f>\
^j
^
a'*
|^(V^I^T^A.)|
^iF^{air(^^^^'+^^^) +
a As
dz
_
"
ei
a<!>_
hi dUi
62
a*_
e3
d<t>
ha r)U3
h2 dU2
[Hint: Let V* = aiCi + a2e2 + asCs and use the fact that
rectangular and the curvilinear coordinates.]
89.
A =
where
88.
grad *,
a*
-ei
(c)
87.
flr
dUi
(a)
Ans.
86.
d(x, y, z)
d{ui, U%, Ma)
Prove that
d*
= V*
dr
6.
in both
CHAP.
VECTORS
7]
MISCELLANEOUS PROBLEMS
If A is a differentiable function
90.
u and
of
91.
Prove formulas
92.
If p and
are polar coordinates
satisfies Laplace's equation.
93.
If
94.
y =
2 cos
6,
and 8 on Page
sin='
1,
is
perpendicular to A.
141.
U =
e cos
^^
^^y
^^^^
6 sin g cos
(a)
(c)
95.
|A(m)|
I59
-5
(4
+B
sin n<p)
sin'e)
(6)
p,
r,
A + B
(c)
In (esc
where
cote)
and
Let T and N denote respectively the unit tangent vector and unit principal normal vector
to a space
curve r = r(M), where r(M) is assumed differentiable. Define a vector B = T X N called
the unit
hinormal vector to the space curve. Prove that
dT
_
-
-d^
dB
'^'
dN
^
^,
= --N,
dF
= .B-T
These are called the Frenet-Serret formulas and are of fundamental importance
in differential
geometry^ In these formulas k is called the curvature, t is called the torsion; and
the reciprocals of
these, p = 1//C and a - l/r, are called the radius of curvature and radius
of torsion respectively.
96.
Prove that the radius of curvature at any point of the plane curve
(a)
differentiable, is given
by
...
(1
97.
spherical coordinates
(r
r'e^
sinx, z
0.
where
f{x) is
of the curve
(57/2, 1, 0)
is
given respectively in
(a) cylindrical,
by
(re
(p
2r|
pi>'')^p
98.
+ y'T'
y"
= f(x),
,,,,,
(p0
2p0)e^
e^
26^
(2r0 sin
2r#0 cos
r0 sin e)e^
respectively.
p, 0, z, r, e,
Let E and H be two vectors assumed to have continuous partial derivatives (of second
order at least)
with respect to position and time. Suppose further that E and H satisfy the equations
V-E =
prove that
E and
0,
V-H =
VXE
V ^
I/,
0,
Cl
VXH =
1 5H
-i|a,
=
.-r
. -
aE
^fl
(i)
Use the
relations in
j^{l(E'
100.
m}
+ cV.(EXH)
An
where i
L,..
hnA[
+ hA2 +
hnAs
M=
1, 2,
[CHAP.
VECTORS
160
101. If
is
i.e.
A,
is
an invariant
(often
Mi
Mi
4-
dy'
dx'
Mi
Mi + Mi + Mi
dz
dy
dx
dz'
The resvdts of this and the preceding problem express an obvious requirement that physical quantities
must not depend on coordinate systems in which they are observed. Such ideas when generalized lead
to
102.
Prove that
(a)
B,
(6)
A X B,
V XA
()%fff?
Vk..V.XV3
105.
in
(6)
(^
)(V..
V. X V3)
terms of Jacobians.
to vectors to
prove relation
(S)
on Page 138.
vectors
a geometrical interpretation of
106.
(c)
is
this.
complex number can be defined as an ordered pair (a, b) of real numbers a and 6 subject to certain
rules
rules of operation for addition and multiplication, (a) What are these rules? (6) How can the
considered
be
can
numbers
complex
why
Explain
(c)
in {a) be used to define subtraction and division?
as two-dimensional vectors, (d) Describe similarities and differences between various operations involving complex numbers and the vectors considered in this chapter.
chapter 8
Applications of Partial Derivatives
APPLICATIONS
1.
to
Tangent Plane
GEOMETRY
to a Surface.
Let F{x,y,z) =
be the equation of a
surface S such as shown in Fig. 8-1. We
shall assume that F is continuously differentiable unless otherwise indicated.
Suppose
we wish to find the equation of a tangent
plane to S at the point P{xo,yo,Zo). A vector
K.
s-
VFl,
normal to S at
subscript
If ro
spectively
r-ro
(r-ro)
is
V^|p =
(1)
Fig. 8-1
perpendicular to No.
In rectangular form this is
is
dF
dx
{x~Xo)
dF
dy
(y-yo)
dF
dz
~ Zo)
{z
(2)
(r
In rectangular
form
- ro) X
No
(r-
ro)
XVF\p
Xa
dF
dx
yo
{3)
Zo
dF
dF
dy
dz
this is
surface
H)
,
161
x,
y and
of the
[CHAP.
162
3.
where u uo.
R=
If
gent to
If ro
f{u)i
+ g(u)i + h(u)k,
at the point
a vector tan-
P is given by
To
^j^.
and r denote the vectors drawn respecto P{xo, Vo, zo) and Q{x, y, z) on
from
tively
- ro
collinear
is
(r-ro)xTo = (r-ro)X-^^ =
In rectangular form this becomes
and G{x, y, z) =
xo
Va
Zo
(6)
F{x,y,
given as the intersection of two surfaces with equations
are
line
the corresponding equations of the tangent
curve
is
Fig. 8-2
(5)
is
0,
F.
G.
Fy
Gy
F.
G,
z-Z(,_
y-yo
x-xa
z)
n
G.
Fy
Gy
F,
G.
4.
of Fig. 8-2
(i.e.
line to
it
(r
- ro)
dn
To
to curve C at Pixo,yo,zo)
at this point). Letting r
follows that
r-ro
is
perpen-
(8)
("^-^"^-d^
when
5.
g'{uo){y-yo)
F.
G.
Fy
Gy
when
the curve
is
{x
xo) +
defined
F.
G.
by F{x, y,
z)
F.
Gx
0,
= f(u),
h'{uo){z-Zo)
(y-yo)
G{x, y,
z)
= g{u),
F.
Gx
+
=
(9)
- h{u), and
F
Gy
[z
-zo)
[10)
0.
Envelopes.
If 4,{x, y,a)
a curve
which
is
be
a one parameter family of curves in the xy plane, there may
that
such
and
family
the
of
member
tangent at each point to some
is
^^^^-
and
^J
" is called
163
JS?/T
DIRECTIONAL DERIVATIVES
Suppose Fix,y,z)
is
defined at a point
4f =
lim ZCa^
lim
.
4s-o As
'l^lTfLnT' '"'
+ Ay,z + Az) ~
dF ^
dFdx
ds
dx ds
"^
dFdy
"^
F(x, y, z)
As^
'^"'"'^''^ ^'
+ AX,
4,^0
'""''""'
ix,y,z)
{13)
^1-^
t*>e
curve
sFdz
^ dzd^
dy ds
T of
(U)
of
DIFFERENTIATION UNDER
^^*
where
Ut
the
^W =
and
tia
may depend
INTEGRAL SIGN
f{x.a)dx
on the parameter
'^
derivatives for
In case
(see
Proll'emfifs^g)""''
INTEGRATION UNDER
if
the
^^
/,.x
'''*'
^:;
z.,
last
"^"'"''''^ '"^^'
gag 6
Then
a.
r^"'
Xe Sinf i!<:^<':f fi'^T
rf
and
= - ^"^
aZ^
'
' ''
z.,
two terms of
^"^"
"^^^"^
-^1-*-^' ^e^-te
integrals
INTEGRAL SIGN
..iitvi;i"rtht'?/ir:;;dttrczsr'
'-
^ "^
'-
-'- '-'"-
The
[CHAP.
164
result is
known
under the
as interchange of the order of integration or integration
integral sign.
dx
If ixo,yo) is
a point
by
fined
dy
'
1.
maximum
2.
(a;o,
^0) is
a relative
minimum
point
point
A>
if
and
and
A>
if
METHOD
A
of
is
obtained
A=
if
relative
^^^
d'f
^
^
>
maximum
<l>{x,y,z) =
or
0,
\
j
or
<
^^^ ^^^
>
(a:o,!/o>
point
MAXIMA
for
de-
32/
[ov
<
^^^
if
A<0.
If
(in
LAGRANGE MULTIPLIERS
^
^
minimum
a relative maximum or
a saddle point.
called
sometimes
{xo,yo) is neither
no information
is
{20)
32/
a relative
4.
"^^
(xo, 1/0) is
is
if
_ jfs2f\(s^\ _ fj^
Kdxdy
dx^jKdy^J
then
3.
and
""
'
and
A<0,
is
(aJo.l/o)
necessary).
MINIMA
values of a function
the formation of the
of
consists
minimum
{21)
X^{x,y,z)
"
dx
"'
dy
'
dz
The parameter
minimum.
which are necessary conditions for a relative maximum or
multiplier.
Lagrange
a
called
is
x,y,z,
of
which is independent
relative
The method can be generalized. If we wish to find the
minimum
^^{xu
values of a function
.,xn)
0,
4>M' ..,^n)
F{xX2,Xs,
...,
0,
G{x,,X2,...,Xu)
maximum
X,
or
Ki^u
F + X,<j., +
.
X,4>,
A,c/>,
{23)
where
A,, A,,
dX2
'
'
dXn
Lagrange multipliers.
...,X which are independent of x^x^, ...,Xn, are the
APPLICATIONS
etc.,
'
to
ERRORS
function of x,y,z,
The theory of diflferentials can be applied to obtain errors in a
28.
when the errors in x,y,z, etc., are known. See Problem
CHAP.
8]
165
Solved Problems
TANGENT PLANE
1.
NORMAL LINE
and
(a)
SURFACE
to a
normal
(&)
+ 3y^ =
2xz^ Sz
(a)
F =
x^yz
+ Sy^ 2x2^ + 8z =
0.
normal
surface at
to the
is
No
=
=
VF|_,,_
Referring to Fig.
Page
8-1,
{2xyz-2z')i
-6i
llj
+
+
(x'z
+ 6y)j +
(x'y
- 4xz + S)kl^_^^_,^
14k
161:
to
The vector
ro =
any point
(x
1)1 +
{x, y, z)
2)j +
(j/
(z
+ l)k
is
ro
+ zk.
= i-|-2j k.
a;i
j/j
lies in
is
thus
perpendicular to No.
- ro)
(r
No
-6(a;
(6)
Let r
a;i
=
- 1) +
is
+ j/j-|-zk
l)k
is
normal
- l)i + (y - 2)j + (z +
or
+ 1) =
O
to
- ro) X
is
{-6i
llj/
+ llj + 14k} =
14z + 2 =
No.
(a
l)i +
(2/
2)j +
Then
No
x-1 y-2
i.e.
-6
which
6a;
ro
is
(r
l)k}
14(z
{(x
i.e.
- 2) +
11(2/
11
+1
14
ll(a;-l)
-6(2/
-2),
-2)
_
~
11
14(2/
14(a;-l)
11(^+1),
-6(z
+ 1)
-6
parameter
2 _
~
14
By
line.
we have
(,
= l-6f,
a;
2/
lit,
14t
2.
In what point does the normal line of Problem 1(6) meet the plane x
Substituting the parametric equations of Problem 1(6), we have
Then
3,
Show
a;
- 6t =
7,
+
2
of surfaces
x^
{2
3(2
llt
lit)
2(14t-l)
-9, z
2yz + y^ =
4a)y^ + az^
14t
10
or
- 1 = -15
+ 3y 2z =
= -1
is (7,
x'
y^
2a;i
(32/'-2z)j
2yz
and
G =
a;^
(2
- 4a)2/^ -
az"
Then
VF =
-9, -15).
F =
10?
22/k,
VG =
2a;i
2(2
- 4a)2/j -
2a2k
'
'
to the
two surfaces at
Ni
all a,
Ni Na
Since
4.
[CHAP. 8
IQQ
(2)(2)
2i
(1,
N2
2k,
0,
2i
- 4a)j -
2(2
4ak
it
where
is given in spherical coordinates by F{r, 6, <l>) = 0,
the
tangent
for
equation
an
(a)
Find
differentiable.
continuously
we
plane to the surface at the point {r^, e^.^o). (6) Find an equation for the tangent plane
(c) Find a set of equations
at the point (2^2, 7r/4, SttM).
to the surface r = 4 cos
The equation
of a surface
suppose that
is
The gradient of *
13*
,13*
dui
hi dU2
All
Pages 141,
3r
hs dUs
142).
m = r,
zk
dUs
1
es
hi dUi
hi 3ui
In spherical coordinates
r sin e cos <pi + r sin e sin 0]
fls
62
ei
,13*
3r
3r
where
(see
(b)
+r
U2
= e,
u%
0,
fei
fe
1,
= r,
As
and
r sin e
xi
+ j/j +
cos ek.
Then
= sin e cos
+ sin 9
= cos e cos i + cos 9
= sin i + cos j
1-61
J
<t>\
62
t 63
sin
sin
cos
sin 9
k
k
(1)
and
,^iaF
3F
As on Page
Now
substituting
VF|p
0.
we have
(i) in (2),
dF
(r-ro)
is
BF
_sin9cos0
+
.
dr
dF
cos
^^g^
f aF\
So
sin 00
r
-~
cos 00
ro sin e
dF\
3F
3<f,
cos 00
dF
TT
dF
-^f
Ai + Bj + Ck,
Denoting the expressions in braces by A, B, C respectively so that VF|p =
This can be written
x) + B{y - 1/0) + C(z - Zo) = 0.
see that the required equation is A{x
for x, y and z in these coordmates.
in spherical coordinates by using the transformation equations
we
(6)
We
have
F -
4 cos
= 7r/4,
0.
Then
= 3!r/4, we
dF/dr
1,
dF/de
4 sin
VF|p
9,
Ai
dF/d-f,
Bj
0.
+ Ck =
-i
Since
u = 2\/2,
From
the transformation equations the given point has rectangular coordinates (- V2, V2,
and so
r-n
{x
+ V2)i
+ {y ~ ^/2)i +
(a),
j.
2),
(z- 2)k.
- + Vi) + - \f2) =
r sin 9 cos = 2V2.
= 4 cos 9 becomes
+
(a;
(3/
or
2V2.
= 4 and
a;^
j/^ + (2 - 2)'
In rectangular coordinates the equation r
cases, however, it may
the tangent plane can be determined from this as in Problem 1. In other
method of part (a)
not be so easy to obtain the equation in rectangular form, and in such cases the
is simpler to use.
CHAP.
8]
The equations
(c)
-110
+ y/2.
X + V2
5.
(3
sin 2t)j
To
III/
lt
(1
sinf)i
to the point
to
equation
- ro =
normal plane
where t - ^tt.
- ro) X
To
- l)k
(z
(z
3j
k.
xi
R = (t
= ^ir is
21
2j
cost,
cos
()i
3k
+ yj + zk.
is
ai
2-2
2/
-^ = ^ = ^1
^
or in parametric form x
2t
+ ir
2
+ j/j + zk
be the vector from O to any point (a;, y, z) of the normal plane. The vector
where t = \w\s, ro = \tt\ + 3j + k. The vector r - ro = {x~ \Tr)\ + (?/ - 3)3 +
the normal plane and hence is perpendicular to To. Then the required equation is
to the point
l)k lies in
(r-r)-To
6.
line is
^n-i
^Jr
0,
ro
is
= t-
to the curve x
3sin3fk|,=i4:,
on the tangent
{x, y, z)
(y - 3)j +
- lv)i +
(x
+ 2cos2tj -
where t=lir
any point
Let r
0,
(b)
from O
the
is
(r
(6)
2/
z-2. Thus
CURVE
to a
at the point
= 1^7j-
a;
the plane
lies in
cos St)k.
-n-
or
cos St
origin
(1
dR
(a)
'
NORMAL PLANE
and
y-yfi
ZZx
y-V2 ^ z-2
member being
TANGENT LINE
167
or
2(a;
-i-,r)
2(1/
- 3) +
3(z
- 1) =
0.
Find equations for the (a) tangent line and (&) normal plane to the curve
= -2, 2xz - x^y ^ 3 at the point (1, -1, 1).
Zx^y
y^z
The equations
(a)
F =
The normals
N2
As
To
in
Problem
The
Page
162.
3{x
5(6) the
and
G ^
0,
2xz
x^y
(3x^
P is
(-61 + j + k) X
tangent line
z
= 2
{(x
or
(6)
1)1
or
normal plane
- 1) + U{y + 1) +
results in (a)
(2z~2xy)i
or
(r-ro)-To =
i.e.,
6xyi
5(a), the
xl_y+l
- ^Q"3
^^'
(6)
- r) X
2/^z
to the curve at
Nx X N2
Thus, as in Problem
(r
=
=
VF\p
VG|p
=
=
Ni
Bx^y
is
+
a;
is
(41
- j + 2k)
31
16j
2k
given by
(y
l)j
+
3i,
(z
l)k}
16t
{31
1,
16j
z
+ 2k} =
=
2i
given by
{(x
2(z
- 1)
1)1
(7)
and
(10)
respectively on
168
7.
[CHAP. 8
is
defined
z)
0,
where we assume
The normals to each surface at point P are given respectively by Ni = VF\p and N2 = VG|p.
Then a tangent vector to the curve at P is To = Ni X N2 = Vif'|p X VG|p. Thus the equation of the
normal plane is (r ro) To = 0. Now
To
juation
V/^|p
Fyj
Fzk)
(G.i
Gyj
Fz F,
Gz G,
Gzk)}
|p
F,
Gx Gy
is
Fy F,
Fy Fz
Gy G.
Fy Fz
G. Gy G.
- ro)
{{FA
F:z
(r
VF|p X VG|p
{x
""'
ajo)
Gy Gz
Fz Fx
(y-yo)
Gz Gx
Fx Fy
(z
Gx Gy
-zo)
ENVELOPES
8.
<^
y, a)
<j){x,
and
of the envelope to be
<^^
f(a),
9(a).
The
slope of
0x
The
slope
""" "^"^^
<t>{x,
{x, y)
of the
envelope ^
at v^,
is
""^ ci.,civ,i/c
"^
/ io
= Q at {x, y) is
_ dy/da _ g'(a)
_
y,a)
dy
^^
dxida
envelope and a member of the family are tangent, we must have
(t>y'
Comparing
(a)
(2)
with
(1)
we
0x/'()
/'(a)
0,
see that 00
56,
dy
dx
can be obtained by
if it exists,
0,
= 0.
given by 0^ dx
Then
at
<i>
dy
or
f'(a)'
0 -'()
(2)
x sin a
+ y cos a =
1.
(6)
geometrically.
(o)
(6)
1.
10.
+ y' =
+ y'^ =
1.
Fig. 8-3
is
From
(2)
x/y. Then
2ax
a^y
substitution in
2
(1)
and
yields x^
(2)
2x
2ay
obtained by solving
CHAP.
11.
8]
of the family
F = 0,
F ^
Then p
x, a
ax
= y and we
i^
I3y
have
F(x, y,
= 0,
F^
z, a, /3)
+ aP =
z = xy.
ax
+ py ap.
/?
Now
43).
F = -x +
0,
if it exists, is
0,
Problem
(see
169
/3
= -y +
F^
0,
DIRECTIONAL DERIVATIVES
12.
corresponding to m
VF =
A
is (1, 1,
2xyzH
du^^"'^
-e""i
"*"
Sx^yz^s.
^*""
^^
cosMJ
"*"
^^^
1,
3k at
-i
2j
(1
in this direction is
To
cosM)k}
sinM)k
~'
k atP
Directional derivative
13.
-2i
("
Then
Since this
2 sinw
Then
1).
x^z^j
e~",
IS
du
is
positive,
is
VF'To
(-21
- j + 3k)
(^
~'
+ ^i + '^ \
_A
i^
Prove that the greatest rate of change of F, i.e. the maximum directional derivative,
takes place in the direction of, and has the magnitude of, the vector V^-
~ ^^*ds
'ds
when
'^
projection of
VF
VF
This projection
direction.
is
~.
in the direction
is
maximum
\VF\.
14. (a)
(a)
VC7
ex^yi
{2x^
- 6yz)i -
to
+ 14] = (3 - 1)1 +
= T =
Sy'k
to
121
i.e.
is
P =
(121
(-1
21
- 2)j +
- 3j +
[5
- (-l)]k =
6k
VW+TW+W
Then
Directional derivative at
12k at P.
14j
- 12k)
21
21
3j
6k.
- 3j + 6k
'^
^i
^^
= _ 90
(6)
From Problem
(")
F rom Problem
1 3,
the
22.
value of the
maximum
maximum
directional derivative is
|121
14j
- 12k.
+ 14j 12k|
170
DIFFERENTIATION UNDER
15.
sign.
Then
f(x, a) dx.
f(x, a
+ ^a)dx -
f(x, a)
+ Aa) da; +
+ ^a)dx +
f(x, a
/(a, a
X2<a) /(,
*^2(a)
dx
a)
^i< + Aa)
^2<a + Aa)
X2(a>
[f(x,
+ Aa) -
the
f(x, a
jCa)
By
+ ha)dx -
f{x, a
+ Aa) d
-^niCa)
^2<a)
+ Aa) diB
-^"ifa)
i<a+Aa)
dx
-',()
-i(a + Aa)
INTEGRAL SIGN
the
/(a;,
[CHAP.
we have
for integrals,
^2<'
,2<a'
+ Aa) -
[/(a;,
/(C, a)]
dx
Aa
/ {x,
|) da;
(1)
^"i (a)
(a + Aa)
Xwj
f{x, a
+ Aa) dir
/(^j,
+ Aa)[Wi (a + Aa) -
-Ui (a)]
(2)
f(x,a
+ ^a)dx
/(^2,
+ Aa)[M2(a + Aa) -
W2(a)]
()
ta + Aa)
X"2
-2 (a)
where
(,
is
+ Aa,
between a and a
^i is
+ Aa)
and
C2 is
between
M2(a)
Then
^
Taking the
limit as
derivatives,
we
Aa
^(a)
By
^) ''^
/ ('''
/(-
^)
^-
^(^- "
^yA<^,a)dx + f[u.{a),a]^-
= f'^^^^dx,
find
f[u,(a),a]^
where ^0.
i>'{a)
Leibnitz's rule,
/
_=..
2\
J
2,
;
<
sin (g
cos
^
r_^
-
By
cos
=
a;
^(a)
X'^IT^:^
2 sin a'
>1
=
a \) J
d
da
()
sin a^
sin a'
find
= (^^^T)^
ft
3 sin a
,
cos a;)^
Jo (2- tl^,2
f'^^ ^
COS
W/
2 sin a"
i/o
'T'^-
2 sin a
L,
_^^
-/^alTi'
Leibnitz's rule, if
aa; da;
fit
sinaa; ["
Jo
obtain
^a^
17. If
^>
making use
-* 0,
g
16. If
'"'
J(a^-l)-'/^
from which
(See
then
f(2-z^
Problem
62,
Chapter
5.)
CHAP.
8]
INTEGRATION UNDER
18.
By
1)1
tp(a)
(1)
/(^>a)
da fdx
Leibnitz's rule,
'P'M
Then by
<f>{a)
from
(1),
Putting a
= 6,
we have
Jj
,p{a)
(2)
f{x,a)daldx
s^ij
integration,
Since
19.
INTEGRAL SIGN
the
Page
{18),
171
0(a)
da
f{x,a)dxidx
<p(a)
Thus from
in (2).
f(x,a)dx
(1)
and
(2)
with c
'/(,) d*
= 0, we
find
f U'-^^^)dx
\a-cosxJ
Prove that
Jo
From Problem
62,
Chapter
'
^^
5,
J^a ++ 'fzl
]
V^^F^J
cos X
'^
-J a'
>
a,b>
ii
1.
1.
rlr^i'^^
riy &-eosa, V
In (a -COS..) 'do;
MAXIMA
20.
TT
In (a
1)
'b
TT
ln(
+ Vb^
)
follows.
MINIMA
and
minimum)
+ Va -
at {xo,yo)
that
is
f{x, y) to
fx{xo,yo)
0,
fy{xo,yo)
0.
be an extreme value for f(x, y), then it must be an extreme value for both f(x, y)
But a necessary condition that these have extreme values at a; = ajo and 2/ = j/o respectively
If f(xo, yo) is to
and
is
21.
Let
in
f{xt y).
fx{xo,yo)
f{x, y)
0, f,{x:,,yo)
some region
fixo.yo) is
fxx{xo,yo)
By
<
0.
mean
(see
+ h,yo + k) -
Page
109),
f{xo, yo)
using fAxo,yo)
f(xo
+ h,yo + k) -
(1)
we
/(xo, yo)
0, /,(a;o,3/o)
^{h^U + 2hkf^ +
Xo
+ eh,
yo
0,
we have
k'fyy)
+ ek
where
(1)
< 9 < 1. On
com-
find
(2)
172
Now
in braces
small h and
all sufficiently
we can
Similarly
22.
+ h,y(, + k) s
But
k.
2/o)
Sx^ S
f(xo,
by hypothesis there is a neighborhood of (xo, 2/o) such that /xx < 0. Also the sum of the terms
by hypothesis. Thus it follows that
must be positive, since f^^Jyy fly >
/(xo
for
[CHAP.
when x
1,
fy
f{x, y)
Sy" 12 =
x^
maximum.
minimum.
when y = 2.
Then
/x.
At
6a;,
A>0
P{1, 2),
and
At Q( 1, 2), A <
At
R(l, 2),
occurring at
23.
fly
is
a relative
hence
0;
is
neither a relative
is
neither a relative
and
<
/xx (or/jj)
so
maximum
maximum
is
36xy.
at
point.
or
minimum
or
minimum
point.
maximum
point.
a relative
minimum
is 2,
point.
maximum
(1)
Volume
(2)
Surface area
or, since z
32/ xy
from
^ = y-^
Dividing equations
a;
then
= V = xyz = 32
of box S xy +
2yz
'.^-
2xz
x^+Mx
= = 4, A =
2/
64
"
when
(S)
J'-
^
^
(1),
y^
of box
For
8-4),
value
What must
If X, y
and
(i),
S..Syy
4 ft
Fig.8.4
(3)
we
Sly
x'y
find
= M,
y
s,o
that x^
(^Y^V^
x-^
= 64
>
LAGRANGE MULTIPLIERS
24.
ffyy
A>0
is 38.
>
and
relative
A =
Then
0.
and Q
A<
At S{l,2),
Thus the
/^
6j/,
for
MAXIMA
and
or x
when
(.4)
= y A and
^^^ Sxx
Xi/^
= 64
= 2.
^>
0.
Hence
it fol-
surface.
MINIMA
Prove that a
Consider F{x,y,z) subject to the constraint condition G{x,y,z)-(i.
FyGx = 0.
that
Gy
extreme
value
is
Fx
F{x,
z)
have
an
necessary condition that
y,
G{x,y,z) 0, we can consider z as a function of x and y, say z f{x,y). A necessary
F[x,y, f{x,y)] have an extreme value is that the partial derivatives with respect to
X and y be zero. This gives
Since
condition that
{1)
Since G(x, y,
z)
0,
we
also
(3)
From
(1)
and
Then from
(3)
(5)
The above
we have
(6) we
and
(5)
find
Px
F=Z.
(2)
Fy
-(-
F^Zy
G.Zx
Gy
G.Zy
(3)
and
have
Gx
^ 0.
(4)
we have
(6)
FyG^-F^Gy =
0.
^^^^-
^J
yields"
where^^/.
26.
"
/:G,-irfG^=^;."
= o!'?i'o'
"
'^'^'^'
+ ^^" =
^='
"
"' '*""
173
we can
consider equivalently
= xF + G
^"
i.x.
+ y^).
=
-
*'
^,
Prom
{!)
and
(2),
8x
since
82/
142/
(a;, 2/)
+1
+
+
=
=
2X0;
2Xj/
or
{1)
(x
or
{2)
4a;
+ l)x + % =
+ (X + 7)j/ =
we must have
(0,0),
^o^Uich
+
+
2a;
+7
:o' Jiriu!,:^eits^
'^-
"'
= "''
^"'
^'
+ 8X-9 =
^"'^^^^^*^''" ^"
^^
or
+ 8.. + 7.^ =
1,-9
225
yieMs -5.^
= 225,
5.
27. (a)
We^mu^t
5+25
find the
1-0
and
and
- ^T+T
+ ^.
(a).
F =
extrema of
^,
+ j/-^ =
.^
o.
+ ^^ + ,.
+ v^ + t?
<?nhiPf.+ tr.
+i,.
t^^/.^e
In this case
,,
multipliers X, X,
2.+^4-X.
0,
x, y, z,
G,
we
find
= ^2X2_
Xi + 4
2^+x.
= O.a =
-5X2
2Xi
10
'
Prom
_^_ + ^i__
+4
2X1
+
Xi
2(X,
17X?
from which
X^
= -10
+ 4)(X, +
245X,
or -75/17.
5)(X.
750
+ 25)
_25_
10 "^2X1
50"
_
-
2.
(X,
%-x.
we
J
Ld
find
(.)
25X2
2X:
+ 50
(^)
,
havet
a'-O, 2/-O,
T" ;j-0 which
wh" I"would
'lr1
not
j-
^^+.
S0
'^
-.
^
.constraint
(M
Pi,"
(&) Give a geometric mterpreta-
10)(17X,
75)
174
Cose
(2),
^X^,
yields X|
1,
^X^' ^
Substituting in the
f^a-
\=
or
180/19
6\/57l9.
The value
2/V5
constraint condition,
two
a;^/4
+ y'/5 +
critical points
+ 45 + 126)/19 =
(20
is
10.
\ = -75/17.
2:
From
+ y^ + z^
of x'
first
Case
\ = -10.
1:
From
2725
[CHAP.
a;=^X2>
(2),
z^/25
2/
1,
of
x''
+ y^ + z^
corresponding to these
(1600
is
+ 1225 + 25)/646 =
minimum
10 and the
is
a;V4
The value
in
yields Xj
value
is
75/17.
75/17.
+ y^ + z^ represents the square of the distance of (x, y, z) from the origin (0, 0, 0), the
equivalent to determining the largest and smallest distances from the origin to the curve
of intersection of the ellipsoid a;V4 + 2/^/5 + z^/25 = 1 and the plane z x + y. Since this curve is
Since
(6)
problem
x''
is
-v/lO
The fact that the maximum and minimum values happen to be given by Xj in both Case 1 and
It follows, in fact, on multiplying equations (!) by x, y and z in
is more than a coincidence.
Case 2
2x'
^+
+
X'
i.e.
APPLICATIONS
28.
The period T
(a)
if
(o)
error and
y^
\,x
obtain
^^ +
22/^
+ K\j
^\i +
z'
conditions,
we
X,2/
2z'
+ y^
-\-
Problem
X,
K(x + y-z)
5 +25^*
^+i^J+
x'
find
= Xj.
z^
76.
ERRORS
to
of a simple
(b)
percent error
T =
we then
Then
2-n-l"'9-'''\
= (2^g-mU-'"dl) +
dT
= -j^
{2^l'"){-U-"'dg)
dl
tt
Error
The error
have from
Error
in
in
is
in
= Ag = dg =
+0.2; error in
is in this
r ^ dT = -=^=^
{-0.05)
ir-x
= Al = dl = 0.05
= 2,
sr
-D
* error (or
/
Percent
relative error)
Another method:
Since
dT _
T ~
Note that
In
1 d;
2
(2)
L^(+0.2) =
= 32
as before.
(1)
Thus we
(1),
V(2)(32)
ih\
(o)
dg
^fl
9^
ylg
T=
is
1.571
is
mi
.
rp
dT
-=r
1
2jr
aI^ "= | =
- 0.025 =
2Y
^^'^^ ^'=
(approx.)
_
-^
0.025
rm l-567o.
1.0/1
-ir/128
ts
vli-
T = In 2s- + ^ In J 1
Idg _ l^ -0.05 \
--^
'-'^^
(*^^)
Inff,
l/+0.2\
2\-r-)-2\-w)
_
-
^
-^^^^''
can be written
Percent error in
T =
|-
Percent error in
Percent error in g
CHAP.
8]
175
MISCELLANEOUS PROBLEMS
29.
Evaluate
x^
r ^x~l
dx.
In
Then by
we
Leibnitz's rule
'<'
X'i(e)-
a,
4>{a)
x'^-
In (a
1)
In 2.
-~
X'
But
c.
+ 1).
In (a
Define
is <p{l)
x" dx
since 0(0)
= 0,
30.
aj
<
finite
>
4,(0)
= (x" l)/ln x,
s ^ 1 and all
a;
0.
b for
which
F{a,b)
IS
and so
The
<
{sin
{ax^
+ bx)y dx
aF/db
a minimum.
The necessary conditions for a minimum are dF/da
tiations,
we
3F
From
aa
56
these
0,
we
C^
)
a^<-sin:
(a;^
Hj^sina;
(ax^+6a;)}2da;
+ MFrfa; =
-2
-2
Performing these
0.
diflferen-
C'^
I
x""
{%yn.x
{ax"
hx))
dx
a;
{sin
(aa;^
6a;)}
dx
a;
find
a;"
da;
a;^
da:
-5IT-^a
6,
we
a;^
da;
a;^
da;
.^0
a;^
a;
sin x
sin
a;
dx
da;
J(,
- -'-4
77'6
find
20
320
-0.40065,
~-^77-*
obtain
b) is
indeed a
1.24798
Tr"
minimum
The polynomial ax^ + bx is said to be a least square approximation of sin x over the
interval (0, jr).
ideas involved here are of importance in many branches of
mathematics and their applications.
176
[CHAP. 8
Supplementary Problems
TANGENT PLANE
NORMAL LINE
and
(2,-4,5).
32.
(a)
x-%y-z =
(a)
SURFACE
to a
(6)
^^
normal
(6)
x^
+ y^
20
/|p(!-a;o)
and
/!,|p(?/-2/o)
at
Az
^.
If a = /(, y), prove that the equations for the tangent plane and normal line at point
are given respectively by
(a)
33.
of the
(6)
P(a;o,
^^
= I-^ =
.5--^
/1p
/Ip
~J-
Prove that the acute angle y between the 2 axis and the normal to the surface F{x, y,z)
is given by sec y = VFf+Ff+FI/\F;,\.
j/o,
at
zo)
any
point
34.
The equation
of a surface
is
P(P( 0o> ^)
A(x
where
xo
cos
A =
35-
= 2,
Ans.
^,
2/0
B(y
yo) +
C(z
zo) =
and
(a)
^^
P'^.lpSin^o,
(a)
^^
+y+z =
=
(6)
^^
y<t _
B
2o
B =
Fp|pSin 0o
-^.xlpcos
C =
<6o.
irz
p<j>
F,\p
at the point
where
^1^
Ans.
and
p sin ^
NORMAL PLANE
0, where F is continuously
the normal line at the point
(6)
and
The surfaces x
z)
xo _
to a
CURVE
Find the equations of the (a) tangent line and (6) normal plane
y = 4 cos 3t, z = 2 sin 5t at the point where t = v/4.
A.S.
37.
xd) +
= 57/2, z = 1. To
2x try + 2irz =
TANGENT LINE
36.
<t>,
i''p|pcos0o
Use Problem 34
p
differentiable.
(6)
3.-6.-52 =
to
3a!-:^-2z
(6)
6 sin
t,
26V2.
2^,
a;
ENVELOPES
38.
Find the envelope of each of the following families of curves in the xy plane. In each case construct
a graph,
Ans.
39.
40.
(a)
(a) x'
= 4y;
ax
(b)
a",
x
(6)
+y =
1,
-~
a = 1.
x y = 1
Find the envelope of a family of lines having the property that the length intercepted between the
X and y axes is a constant a.
Ans. x"^ + y^'" = a^'^
Find the envelope of the family of circles having centers on the parabola y
vertex. [Hint: Let {a, a^) be any point on the parabola.]
Ans. x^ =
its
41.
~ -J**
through
CHAP.
42.
8]
177
iz
(a)
= (x-y)\
-y) - aH =
= z^^2xz
a(x
(a)
Ans.
y^
(6)
- af +
(6) {x
1,
2/^
2az
43.
Prove that the envelope of the two parameter family of surfaces F(x,
y,
obtained by eliminating a and /? in the equations F = 0, F = 0, F^ = 0.
44.
Find the envelope of the two parameter families (a) z = ax + /iy z cos y = a where cosV + cos'/3 + cos'y = 1 and a is
a constant.
Ans. (a) 4z = ^ + J/^ (6) x^ + y' + z" = a'
a'
z, a,
- p'
p)
and
(6)
0,
a;
if it exists, is
cos a
2/
cos
j8
DIRECTIONAL DERIVATIVES
45.
(a)
(6)
Ans.
46.
U =
what
direction
(a) 10/3,
-2i
(6)
- z'
2xy
+ 4j - 2k,
maximum?
is
2V6
(c)
maximum?
Ans.
12\/3-6;
(a)
47.
Prove that
point
is
4,,
z)
12v5
DIFFERENTIATION UNDER
the
50.
(a)
If
F{a)
51.
52.
53.
Given
^ -I-^,p>_i.
x" dx
Prove that
Prove that
Show
that
-^
by Leibnitz's
rule.
In (1
cos a)
In (1
2 cos
Cj^-i^
(5-3 cos
jlj^
a;
r.
a;)'
cos
4
(a'
x')
77-
x^lnx^dx =
~"
"^
ln(^ ^
a^) da;
in
(a)
1"
7'
l"i
<
H<1
'^
^^^~t, m =
1-
lal
1.
2048
dxXda
sin x)
J |J
dx
{a
(a^
- a;^) da Ida;.
2ira,
xY} dx
all
constants a and
1277
{(6-
by direct
1,2,3.....
^.
J-2V {a -
any
cos a*
INTEGRAL SIGN
the
at
2v^
- 1 In (a^ + 1)
Prove that
da;
(6)
J^
INTEGRATION UNDER
Verify that
-'vzF
^^
find
2a tan"' a
-\
Ans.
-f""
Ans.
54.
find
integration.
J^
INTEGRAL SIGN
^
tan-'|da;,
(a)
directional derivative of
L(^\ + /'MLV
0?
maximum
is
(^\' +
-J
given by
(c)
F(p,
if
(6) in
+ 2j;
direction
sin
xY
sin
2j7(6^
- a^)
6,
D.. a,,
56.
[CHAP.
178
'
result
-^
sec
In (1
^^.
.>1
.. pr.v. that
- -^.
i COS x) da
LAGRANGE MULTIPLIERS
and MINIMA.
MAXIMA
a;
58.
59.
What
a;V9
60.
a;
2/
61.
62.
Prove that
63
(a)
is
ABC
there is a point
the intersection of the medians.
in
any triangle
maximum =
5,
such that
values of f(x,y)
(b) Can the result of
y equal
on the surface
2x'
3x^
6,
>
0,
2/
>
0,
Find the
that
is
=
(a)
+ 4xy + Gy^ =
is
140.
a minimum and
x'
Explam.
to zero.
35.
minimum = 5
65.
values
case where we wish to find the extreme
Establish the method of Lagrange multipliers in the
0.
H(x,y,z)
G{x,y,z)
0,
conditions
of F{x,y,z) subject to the two constraint
66.
67.
APPLICATIONS
68
69.
to
ellipsoid
\\x'
+ 9y^ +
surfaces
to the curve of intersection of the
15z^
- Axy +
lOj/z
- 20x2 =
80.
xyz
=a
Ans. 64,rA/2/3
ERRORS
is
^^^^^^^
computmg
"
^he
15.0 feet, and the included ^"eJ%^0-0sides of a triangle are measured to be 12.0 and
while the angle can be measured to with n 2%
accuracy
1%
within
to
measured
be
can
lengths
side
in determining the {a) area and (6) opposite
find the maximum error and percent error
The
accuracy,
of the triangle.
Ans.
(a)
(6)
0.287
ft,
2.08%
MISCELLANEOUS PROBLEMS
70.
and
4>
CHAP.
71.
8]
Find an equation for the (a) tangent plane and (6) normal line
where r-1, e~ ttIA, = n-/2, (r, e,
being spherical coordinates.
179
jr'
at the point
</,)
Ans.
72.
Ax
(a)
(77-^
477)1/
(47r
-7r'V2,
(6)
(a)
-4
73.
The
potential
where p
is
Prove that
~X-dx
If
<
4ac
and
2n-/V4ac - 6^ [Hint:
ax^ + bxy + cy^ 1.]
76.
the plane
Z--V/272
77^-477
Ax + By + Cz + D =
+D
2x~iy + Qz =
cos
(r,
tf,
.p)
(b)
by
any point
is
4 cos^ e
m > 0,
if
'
Ans.
20.
directional derivative at
g
is
-1?
fc'
'<m)
In
maximum
a constant.
+ 477
is
pVsin"
75.
Cc
T,
74.
77^
(a, b, c) to
Aa + Bb +
(6)
_
~
3/-\/2"/2
- 77^)2 =
>
0,
Find the
c>
0,
Prove that the maximum and minimum distances from the origin to the curve of
intersection defined
by xVa' + yVb' + z'/c' = 1 and Ax + By + Cz =
can be obtained by solving for d the equation
BV
C'c'
77.
Prove that the last equation in the preceding problem always has two real solutions
df and di for any
real non-zero constants a,b,c and any real constants A,B,C (not
all zero).
Discuss the geometrical
significance of this.
78.
(a)
dx
Prove that
(6)
(C)
Find
lim Im.
"-*
79.
80.
81.
(a)
82.
Use
(a)
Find
(b)
Examine
[Hint:
(i.e.
is
+ y^ which
f(x, y)
(x^
cos X dx
is closest to
the point
(3,
-6,
prove that
2x + Ay^ SyY.
2(a^
cos'
(2
cos x
a;
1)
da
sin
a^
877
xY
8 In 2
50
For
w =
(a)
x''
use
4)
In a
a cos X ^ - sin X
'""^
(a) to
of
x^
+ ay
25
p^^
Prove that
Jo
(6)
(x'
da
z
+ aj
dx
A. lim
(x'
(x^
da
M_oo
f JCj~^^
+
lim
Is
J_tan-^+
+ a'^Y
(x^
positive definite) if
A >
0,
A D
D B
>
0,
A D F
D B E
F E C
>
chapter 9
Multiple Integrals
DOUBLE INTEGRALS
of
Let F{x,y) be defined in a closed region
subn
into
Subdivide
9-1).
Fig.
the xy plane (see
w. Let (|^, r,^)
regions A%^ of area aA^, fc = 1, 2,
sum
the
Form
A%^.
be some point of
F(^^,r,^)AAu
k=l
Consider
n-*oo k
limit is taken so that the number % of subdivisions increases without limit and such that the
where the
Jj"
and
is
F{x,y)dA
Fig. 9-1
{3)
%.
F{x,y)
is
if
ITERATED INTEGRALS
of % in at most
such that any lines parallel to the y axis meet the boundary
curves ACB
the
of
equations
the
two points (as is true in Fig. 9-1), then we can write
and
where
respectively,
Mx)
=
/2(^) are
Mx)
and ADB bounding % as y = fi{x) and y
double
the
evaluate
can
we
case
this
In
single-valued and continuous in a^x^b.
grid
ot
a
constructmg
formed
by
rectangles
integral (3) by choosing the regions A%,, as
can
be
Then
areas.
corresponding
{3)
the
lines parallel to the x and y axes and aA^ as
If
is
written
^^
ff
^i>
F{x,y)
dxdy
^h<.x^
F{x,y)dydx
j
' y
(*)
= fi'.x'l
F{x,y)dyydx
i
I
180
^^^^-
MULTIPLE INTEGRALS
9]
If
is
x~gx{v) and x =
g^iy)
respectively and
J J F{x,y)dxdy
we
t Tsed^Ve
into regions
{J,)
and
in at
F{x,y)dxdy
(5) yield
g?
most
find similarly
other fornT
jgj^
(5\
F{x,y)dx\dy
(See,
however Problem 17
^^
''^
^^^ ^ (^)' ^^^^^-^ - approp^ate miy
calloVel^^^^^^^^" ^^^r.;.a^. o/
^^^ order of integration with respect to
the
"w' 1
shown
'V ''whl5f^'
which
<2^j,q^^,
TRIPLE INTEGRALS
The above
rj-;
we form
linear
Hnt'r^'Lnsl^In'f"
dimension of each
"^"'k""""'
subregion
^j F{x,y,z)dV
r^rtc^lfSlfi^lif^! ^- ^-
'''
'^-^*
(7)
^-
-^-
^^
^(--)
continu-
'
region
is subdivided into '"f*'"^
region'risTubd?vMed^[nt
subregions
TRANSFORMATIONS
of
MULTIPLE INTEGRALS
The
[CHAP.
MULTIPLE INTEGRALS
182
we
let {u,v)
Cj" F{x,tj)dxdy
G{u,v)
where
JJ
G{u,v)
d{x, y)
dudv
{9)
d{u, v)
and
F{f{u,v), g{u,v)}
du
dx
dv
dy_
dy_
du
dv
dx
djx, y)
d(u, v)
{10)
du dv dw
{11)
3?.'
where
dy^
du
d{u,v,w)
is
dx
dv
dy
dv
and
dx
dW
dy
{12)
dw
dz_
dz
dz
du
dv
dw
and w.
the Jacobian of x, y and z with respect to u, v
and triple
and {11) correspond to change of variables for double
results
The
{9)
integrals.
made.
Generalizations to higher dimensions are easily
Solved Problems
DOUBLE INTEGRALS
1.
in the
xy plane bounded by
(a)
{b)
(c)
jj
{x^
y^x^
= 2,
y=^l.
+ y"") dx dy.
'R.
(a)
(6)
is
(b).
shown shaded
CHAP.
MULTIPLE INTEGRALS
9]
183
!/
X
x
=l
a;
=2
Fig. 9-2
Method
(c)
1:
The double
Fig. 9-3
dx
j=i
1006
105
ttT? cm V-Tr^-Y
betLen"=7and ^^ri
Method
2:
{x'
+ y')dxdy
f V?
(x'
(^
+ y')
y^}dz\
dx [dy
'"'
colum^ as n
riTto
respect to
pL
S'l^*-
T^'V'-
lb
''^""
xy-
rfj/
=v
""
1006
105
''f''"'^
/="l
an7,-V""'
'"
for
'
Required volume
r ^3
-' = !
co3ns LLeen
^'
9 S
1^1
II
all
x^
such horizontal
+ y^ =
a-
and
nnt
zdydx
'
+ j'
^Vaz-xz
-y
(ffl='-a;^)(;a;
16a^
Iw rarSfshadedfn
-
sucht
volu'mT
,F^
tb
//""""l'
[CHAP.
MULTIPLE INTEGRALS
184
+ y)} dydx
(x
^9 =
x=
in Fig. 9-5
- x)y -
(6
dx
hy'
j,
|(6
- xy dx
36
Fig. 9-5
{6
darkly shaded) corresponds to
In this case the volume of a typical column (shown
of. the
over th^ region
integrating
by
obtained
then
are
integration
of
limits
(x + y)} dydx. The
(obtamed
to y to y from y =
figure. Keeping x constant and integrating with respect
parallel to the yz plane
slab
in
a
columns
all
summing
=
to
=
corresponds
x + y)
from z 6 and z
= to ; = 6 corresponds to adding the volumes of all
Finally, integrating with respect to x from a;
volume.
required
such slabs and gives the
G-x ^
TRANSFORMATION
4.
DOUBLE INTEGRALS
of
jj F{x,y)dxdy.
We
= cops*^
Let
(m, v),
of Fig. 9-6
is
given approximately by
But
du
\^-^
dx dy
du du
Sv
dx
dv
ler
SO that
is
k
-
is
3{x,y)
3(m, v)
d{x,y)
Am Av
\s{u, v)
sum
g{u, v)}
d(x,y)
AuAv
d(u,v)
If
where
dy_
dx dy
dv dv
2 F{f{u, v),
An
dx
Su du
-r-lilUikV
dv\
Am.Ai>.
\du
dy
dv
Fig. 9-6
is
P{i(n,v),g{u,v)}\^^\dudv
is
mapped under
the transformation
y = g{u,v).
^ ,.
^
i
of variables makes use of Ime integrals
Another method of justifying the above method of change
Problem 32).
and Green's theorem in the plane (see Chapter 10,
f{u,v),
CHAP.
li
5.
MULTIPLE INTEGRALS
9]
u~
From
d(u, v)
__
9{x, y)
the identity
{x^
(a;2
Then by Problem
^ yy =
Chapter
{:.2
v}
y)
d{u, v)
y,
- 2/^)^ +
^-
i^2xyY
2a;
-22/
22/
2a;
terms of u and
v.
we have
and
v^
in
a;^
= V^?T^
2/^
6,
^(a;,
Anothe^r .ethod:
d{x,y)/3(u,v)
Ux Uy
+ 2/T =
45,
find
185
d{u, v)/d(x, y)
4(a;2
+ y^)
4Vm'
i;^
xy-i, xy =
1.
Fig. 9-7
R,l-l polar
= u
..,., i.
2a;i/
JJ
%P
4/
7.
Evaluate
a;2
+ 2/2 =
(see
+ y^ =
%'
4v';7T^
,,|&M|,,,
OKU, v)
/-
dw
di;
4j^^J^^^
5.
i.
^^^^^'^^'
t\t ra:2rt:tirtf
*'^^"'*^*'" ***^ ^^'
fr,.. +
I"/
9.
pilCrctt:^;^"Sr
,ere
+ ,,,,,,
J/v'J^..,,,
4 and x^
,..
__dudv_
^^ V^JTV'
=
where we have used the results
of Problem
t>, ,..
r'^-^
[Pi^- 9-8()
os..
below]
2.
is
. sin,
mapped
[CHAP.
MULTIPLE INTEGRALS
186
Itt-
ir~
:m:
1
!I
(6)
Fig. 9-8
S{x,y)
Since
follows that
it
d{p,4.)
S{x,y)\
I
y/x"
dpd</>
+ y^ dx dy
\j p-p dp
d,p
"K
19
d(^
38ff
We
p dp
8.
in Fig. 9-8(a).
is
By
Fig. 9-9
is
O'
laVcos
,77/4
,77/4
pdpd<i,
2c()
d0
2
a?-
cos 20 d0
a^ sin
20
f/V
- dtdy
X77/4
TRIPLE INTEGRALS
9.
(a)
Sketch
the
bounded by x
2/
(5)
= 0,
-I-
dimensional region %.
a; = 0,
1/ + 2 = a (a > 0),
= 0.
\\\
[x^
J-
1/
+ if + z^) dx dy dz
(c)
(a)
(6)
{x, y,
Since ^ + j/^ + z^ is the square of the distance from any point
inertia
the triple integral as representing the polar moment of
is
shown
(b).
in Fig. 9-10
Fig. 9-10
z)
to (0, 0, 0),
(i.e.
moment
we can
consider
of inertia with
CHAP.
MULTIPLE INTEGRALS
9]
We
The
density varies as x^
(c)
187
triple integral
+ y" + z^.
if
the
(x''
y"-
z^)
dz dy dx
x^z
y^z
~x)-
x%a
dy dx
-?^+
rLv-^)2 -
x^a -x)y
+ (a-
x-y
x)y^
i<^~^)y'
+ (^_JL^yY
y^^ ^^
_yl_ (a-x-y)
ia-xY
(-x)^
g!(g^:M!
y^
(a~xV
f f x^a~xr
(a-xa
6
rdx
J
dx
20
a-x-v
all
IS
10.
Find the
volume and
(a)
'
.7''
(b)
= 0,y = 0,y^6,z =
sta"nt
The region
(a)
9^ is
shown
Required volume
in the
in Fig. 9-11.
fffda; dy dz
(4
dzdydx
a;2) dy dy.
*^ X = (i *^y =
(4
(24-6x2)
dx
x^)y
da;
32
Fig. 9-11
oJ,.X=
"d^'^ydx
32. by part
(a),
since
<r
is
constant.
24(7
Total mass
Total mass
32(7
Then
[CHAP.
MULTIPLE INTEGRALS
188
moment about xz
Total
Total
_
-
_
"
32<j
azdzdydx
32a
Total mass
Total mass
960-
Total mass
Total mass
moment about xy
^""plane
ay dz dy dx
(3/4,3,8/5).
because of symmetry.
Note that the value for y could have been predicted
TRANSFORMATION
11.
of
TRIPLE INTEGRALS
Page
182, for
changing variables
in a triple integral.
Fig. 9-12
The vector
"R.
we
4,
==
xi
3/j
to point
zk
is
/(u,i;,'M;)i
fl'(w,-i',w)J
x^f{u,v,w), y
'l(^t,'y,')k
9(u,v,w)
and
h(u,v,w).
to
Tangent vectors to the coordinate curves corresponding
of the region
volume
the
Then
,rle^.
Brl^v,
ar/a,
by
surfals are gWen
approximately by
S{x,y,z)
Au At) Am
A^
\du
The
dv
d{u,v,w)
dw\
An
sum
Am. A-y
Aw
is
du dv dm
where '^
is
uvw
is
CHAP.
12.
MULTIPLE INTEGRALS
9]
Express
F{x, y,
J JJ
dx dy dz
z)
(a)
As
in
Problem
and
d{x,y,z)/d{p,4,,z)
6,
spherical coordinates.
(6)
Chapter
39,
in (a) cylindrical
189
.x
p cos 0,
Then by Problem
p.
becomes
III
p sin 0, z
= z.
G{p,ip,z) pdpd(f,dz
%'
where %'
p
sm 0,
is
p,
(6)
0, z
space corresponding to
r cos
spherical
in
III
13.
sm
is the region in
cos 0, r sin 9 sin 0, r cos
H{r,
e,
%'
the r,e,<p
where %'
(r
and where
G(p, 0, )
coordinates
are
r sin
cos 0,
F(p cos
y-r sin e
0) r^ sin
r^ sin .
dr de
0,
sin
d,
The volume
triple
d<p
space corresponding to
%, and where
H(r,,0)
9).
9.
z).
x^
+ ifl
is
coordinates.
= p' and
Required volume
spectively z
4 times
4 X1T/2
p-a.
Then
volume shown
^a
I
in Fig. 9-13
pdzdpdtp
J-.ir/2
p^
dp
d<t,
0, l' -
or
.=0 4
fr^m
It"
11
2
Fig. 9-13
oZZ
;/'
/)
to z (keeping p and
constant) from ^
p^ corresponds to
to z
by dV) in a vertical column extending from the xy plane to
""t
subsequent
mtegration with respect to p (keeping
constant) from p =
to p
a
"*"' (indicated
-.'"^u
The
ff
Finally, integration
shaped regions.
also be
performed
"^*'''
manJ/bv
"T^^^ transformation.
mapped
by V''r'^
the cylindrical coordinate
^^'
same
result.
moment
?^
1?
thatT'^^
the density
(a)
is
The moment of
which
is
is
^7r/2
p^
4o
p^dpd^
apdz dp
d,p
4<r
d0
va^a
[CHAP.
MULTIPLE INTEGRALS
190
The
Note that
(b)
so that
in setting
The
-a'a
volume X density
p^
p'
JJJ
ap dz dp
- -3- "
^a'
'
by Problem
can think of op dz dp d^
as the
Problem
such that
13,
3^'*'^
moment
the value
is
^-
L we
d,f>,
mass
z axis.
13.
MK' = |Ma%
i.e.
K'
K = a^/2/3.
or
|a^
cylinder would be h.
z^ = a^ and
Find the volume of the region bounded above by the sphere x^ + y^ +
such that O^a^^.
below by the cone z^ sin^a = {x^ + y') cos^a, where is a constant
sphere of radius a.
(b) From the result in (a), find the volume of a
15. (a)
= a.
This
y-r sm e
sm 0,
(a)
(r^ sin^ e
r^ cos" e sin' a
I.e.,
say
r^ cos^ e sin'
= tan
and
so e
=a
or
cos^
=
e =^
<p)
cos' a
ir
- a.
It is sufficient to consider
one of these,
= a.
Required volume
volume (shaded)
4 times
in Fig. 9-14
^a
-,0!
J,TT/2
r' sin e
dr de
d,j>
"O'
=:
i_
^0
"
d^
d4>
sine ded^
de
sin 9
-s-
^ = 0^0 =
V_ dl'
^6 =
4o' X7r/2
3
) = n
cos
e\
dtji
18=0
-(1 -
Fig. 9-14
cos a)
to r- a corconstant) from
(keeping e and
in a column
dV)
indicated
by
(such
as
elements
cubical
responds to summing the volumes of all
constant)
respect to e (keeping
with
integration
=
=
subsequent
The
a.
r
to
r
from
extending
shaped
wedge
the
in
columns
all
of
volumes
the
summing
from 6 = to e = ,7/4 corresponds to
to adding volumes of all such wedge
corresponds
to
respect
with
integration
Finally,
region.
shaped regions.
(6)
Letting
a-w,
r-0
to r
cos
tt)
is
-^va
CHAP.
16. (a)
(b)
{a)
MULTIPLE INTEGRALS
9]
Use the
Problem
in
191
15.
The centroid
{x, y, z) is,
Total
= y=^0
and
fSS
zadV
/J"/ dV
Total mass
Since z
= rcosB
and a
is
'''^^"*'^^'^^*
J^=<,J,=J,_/'=^*
is
^''j
ad'
Then
= b-/2,
P.ove that
(e,
Toa* sin^ a
la.o
by
<r,
is
|r<7a(l
- cos a).
cos a)
"
8"^^
,,,
J^' jj^'
'=^)-
= ^a.
MISCELLANEOUS PROBLEMS
...
fTraa^d
Letting a
sin^g
r
.
^Toa* sin^ a
\Traa*
a:
_
(6)
Sincos9d6id0
f^[f^
.,},.
-'o
V(a;
Jo
(x
1
,
dx\dy = -^.
dx
+ 2/r^a; + 2/y
(6)
lf
results^''
results.
(a) to
[-'
^ -jydxVdy = I
(x
A Z^m^-^
fcondition
A
sufficient
the
region
origin.
18.
1,
O^ySl
The mtegral
Provethat
Let
a;
I(x)
is
In this case
ff^^axdy,
fails to exist
j^ |j^
is
the
where
doubL
F{u)dujdt,
I'(x)
J{x)
F(u)du,
(x-u)F(u)du.
J'(x)
Then
F(u)du
[CHAP.
MULTIPLE INTEGRALS
j^g2
The
The
form
Jr
F(x)dx'-
('
...
54)
(;r^/(^-)""^Wdr*
CF{x)dx'
(x-u)F(u)du
Problem
Supplementary Problems
DOUBLE INTEGRALS
y 19.
in the
v4o.
problem.
Find the centroid of the region in the preceding
/21.
Given
j^{x + y)dxdy.
the doubl"integl-al.
Ans.
22.
(6)
(6)
(c)
Fmdj
to-
= x.
(6)
(x
(a)
Ans. x
^,
1
j
interpretation of
Sketch the region and give a possible physical
+ y)dydx,
(c)
241/60
(c)
Showthat
+ X=J.=V.""^ '
bounded by x/a + y/b + z/c =
""'
X^J^^^-'^g'^^'^^
23.
24.
25.
x'
+ y\
= 0,
= -a,
Evaluate
If
is
of
= a,
= -a,
DOUBLE INTEGRALS
(T V^^T^dxdy,
where
is
JJ e-'''+>'>
+ y' S
dxdy.
a\
Ans.
<^
28.
TRANSFORMATION
27.
^/26.
By using
the transformation
+ y = u,y=uv, show
that
dy dx
x=
*^y~0
e'"'-^'^
-^
Ans. %^a^ v^
,r(l
- e""')
= a.
in
Problem 24
CHAP.
29.
30.
MULTIPLE INTEGRALS
9]
= 4,xy:= 8,
31.
32.
Let
[Hint:
in the first
= 5,
xy'
1,
= Q,
xy"
= 15.
= u,
Let xy
[Hint:
xy'
= v.]
quadrant bounded by
I93
2/
= a;S
= 0. Show
J/
= 4a;S
that
a;
= j/,
a;
= 4j/.
Ans. 4
=
(( co/^^^^\dxdy
\x
"
+ yJ
-'-'
5HlI
2
'
TRIPLE INTEGRALS
^33.
Evaluate
(a)
34.
35.
J J ^^^ xyzdzdydx.
J^
^
Find the
where a,
(a)
h, c
Find the
Ans.
(a)
(a)
volume and
(6)
are positive.
moment
Mia'
of inert ia and
+ b'^j/lO,
(b)
\/(a'
(6)
39.
40.
^'
TRANSFORMATION
x'
+ y'
and
z^4-x'-y'
38,
Evaluate
cone
^^^
y/x'
mtegration
= 2x.
iWe.***^*
42.
43.
Vx'
in
Problem
where
is
JJJ
^-^-^^^___^ ^here
Tni latttiTb)'^
''
45.
is
"'"' a>b>0.ib)
Ans. 8^
= ^^^^+^
- l)/i
^'^'l **>
paraboloid z
6,
about
=^ x'
^^^^^
a
,r/2.
- = 2. cos., and
Ans. ^^a'{l- cos^ a)
+ y^ + g^ =
beW
(a) IS
ia)x = y^O,
|(a^- 6^)/(3_
j,).
+ y'
axis if the
'"1''
!rL?ltro'r;//f
a where 0<a<^/2. %ftfDiscuss
n"
the case
(6)
f ^'iJ^e'^'fevt^^ft'ef^
its
x'
and the
Perform the
[Hint-
(a).
is
=3
Evaluate
the density
if
a.
Ans. ^/2
27.(2v/2
*'
+ y/h + z/c ^
TRIPLE INTEGRALS
of
(a)
by xia
^Xlfox'^r'''Ans'l^'''''''''^"^''^'^^'
38.
(a).
+ b^lO
''
37.
(6)
'
if
a'
(a).
by the cone
the densitv
a=b
(j)
= ^ = o.
[CHAP.
MULTIPLE INTEGRALS
194
MISCELLANEOUS PROBLEMS
46.
47.
48.
49.
b, if
Find the (a) volume and (6) centroid of the region bounded above by the sphere
and below by the plane z = b where a > 6 > 0, assuming constant density.
Ans. (a) l^(2a' - Sa'b + b'); (b) x = y^O, z = |(a + b)V(2a + 6)
x'
+ y' + z' =
a^
A sphere of radius a has a cylindrical hole of radius 6 bored from it, the axis of the cylinder coinciding
6^)''*].
with a diameter of the sphere. Show that the volume of the sphere which remains is j7r[a^ - (a" simple closed curve in a plane is revolved about an axis in the plane which does not intersect the
Prove that the volume generated is equal to the area bounded by the curve multiplied by the
distance traveled by the centroid of the area (Pappus' theorem).
curve.
50.
Use Problem 49
about the x
to find the
circle
51.
52.
Evaluate
x^/a"
f)
x^
+ (y-bY =
a^,
>a>
Ans. 2n-V6
axis.
VT^^TxV^T^WTzV?)
+ y^/b' + zVc^ =
1.
[Hint:
Let x
dxdydz,
= au,
where
bv, z
'5^
is
= cw. Then
xy
= l,
xy=:9, xz
4:,
xz
= 36,
Ans. ^iT^abc
53.
If
'ff
is
[Hint:
the region
Let x
Prove that
+ xy + y^ ^
= w cos a
Then
in the integrand.
54.
x^
-u
let
1,
sin a,
u = ap
C' F{x)dx'^
li
prove that
= m sin a +
cos
<p,
^^w
e-<"^^+^+^'
da;% = -p(e-l).
V3
%
i>
cos a
bp sin
(x
- u)''-^ F{u) du
for
w=
1,2,3,
..
chapter 10
Line Integrals, Surface Integrals
and
LINE INTEGRALS
Let C be a curve
connects
xy plane which
and B{a2, hi), (see
Let P{x,y) and Q{x,y) be single-
points
Fig. 10-1).
Theorems
Integral
in the
A{ai, bi)
2/n-i).
Call AXfc
1,2,
.,n where
(ai, 61)
\yk
(4,
Form
{Xk, yk).
(Xi,y,)
= yk- yk-i,
{xn, yn)
(k+i.ii.+i)
77^,)
are
Ol
at
C between
the
sum
Fig. 10-1
{P{ik'Vk)^^<^
(i)
Q{k'nk)^Vk)
The limit of this sum as tc ^ =0 in such a way that all the quantities A^^, Ayk
zero, if such limit exists, is called a line integral along C
and is denoted by
P{x,y)dx
+ Q{x,y)dy
or
approach
Pdx + Qdy
(2)
The limit does exist if P and Q are continuous (or sectionally continuous) at all points of C.
The value of the integral depends in general on P, Q, the particular curve C, and on the
limits (ai, 61) and (a2, 62).
In an exactly analogous
three dimensional space as
lim
{A, (4,
Q AXk
r,,,
^2(4, rj^,
A?/,
A3(4,
,;
Q AZk}
in
(3)
y and
z.
Other types of
example,
{xk,yk)
if
and
lim
is
195
U(x, y) ds
U)
196
VECTOR NOTATION
for
[CHAP. 10
LINE INTEGRALS
J^
fA
special
If
field is
(dxi
dy\
dzk)
(5)
dr
A =
where
Aii
EVALUATION
of
work done
in
is
force
(^)
C.
LINE INTEGRALS
If the equation of a
{2) is
integral
^"2
P{x,f{x)}dx
,^^
Q{x,f{x)}f'{x)dx
(7)
which
is
Similarly
if
is
given as
a;
g{y),
then dx
g'(y)
dy and the
line integral
becomes
is
C
where h and
ta
^(),
^(t),
+ Qm),m}rit)dt
p{<i>{t),m)4>'ip)dt
corresponding to points
(8)
may
()
and
respectively.
PROPERTIES
of
LINE INTEGRALS
of ordinary integrals.
Line integrals have properties which are analogous to those
For example:
1.
2.
P{x,ij)dx
+ Q{x,y)dy
Pdx + Qdy
J(oi.bi)
Thus reversal
=
-
j" P{x,y)dx
J'^Q{x,y)dy
Pdx + Qdy
^'(aj.bj)
line integral.
of the path of integration changes the sign of the
CHAP.
10]
^(03,63)
J''^2.f'2)
where
Pdx + Qdy
(as, 63) is
^(02,62)
Pdx + Qdy +
197
Pdx + Qdy
another point on C.
MULTIPLY-CONNECTED REGIONS
and
If a plane region
GREEN'S THEOREM
in the
PLANE
Pdx
where
j)^
is
Then
^ Qdy
Xr(g-|V,,
closed
is
and that
it
(,,)
is
direction.
CONDITIONS
Theorem
path
for a
LINE INTEGRAL
to be
INDEPENDENT
closed curves
of the
(i
PATH
1.
dy
where
more
it is
J^Pdx + Qdy
is
that in
to be independent of the
dQ^
dx
(^^)
198
i.e.
Pdx + Qdy is an
Pdx + Qdy =
exact differential,
In such case if
d,j>.
is given by
integral
the value of the line
The condition
holds and
if {11)
and
{x^, 2/2),
(xi, yi)
Pdx + Qdy
In particular
are
is closed,
[CHAP. 10
d4>
we have
<i>ix2,y2)
a;i
= ir2,
1/1
(12)
<l>{xuyi)
= 1/2 and
The
results in
Theorem
Theorem
^ Aidx
+ Azdy + Asdz
to be inde-
Thus we have
Ml
M2
dAs
Ml
dy
dX
dX
dz
'
M2
M3
dz
dy
is
that in
%
{U)
it is
The
11-13.
2.
where
(i5)
terms of vectors.
A-dr
and condition
{1I^)
If
A =
is
Aii
A^.^
Ask,
A =
If
0.
V A
condition
The condition {lA) [or the equivalent condition V X A = 0] is also the
i.e. that there exists a
diiferential,
exact
an
is
A-dr]
[or
Asdz
that Aidx + Aidy +
case if the endpomts
function <i,{x,y,z) such that A,dx + A^dy + A^dz = d^. In such
integral is given by
of curve C are (x^yuZi) and {x2,y2,Z2), the value of the line
In particular
if
A-dr
is
closed
d<j>
<l>{x2,y2,Z2)
^{xuyuZi)
and
A=
0,
i A'dv
we have
{16)
SURFACE INTEGRALS
Let S be a two-sided surface having
on the xy plane as in the
projection
adjoining Fig. 10-2. Assume that an
equation for S is 2 = f{x, y), where / is
single-valued and continuous for all x
and y
in
%.
Divide
of area AAp, p
% into n subregions
1,2,
{15)
.,n,
and
erect
36/-
Fig. 10-2
CHAP.
10]
Let
(j>{x,y,z)
all
points of S.
Form
the
199
sum
p=l
where {^^, rj^, CJ is some point of aS^. If the limit of this sum as % -> in such a way
that each ASp -^
exists, the resulting limit is called the surface integral of ^{x, y, z)
over S and is designated by
JX <j){x,y,z)dS
= [sec y^l AAp approximately, where y^ is the angle between the normal
the positive z axis, the limit of the sum (17) can be written
ASp
Since
line to
(18)
S and
JJ
The quantity
]sec y| is
>j>{x,y,z)\secy\dA
(ig)
given by
z = f{x, y)
has continuous (or sectionally continuous) derivatives
can be written in rectangular form as
%,
{19)
SUi^.y,^)^ i+(|J+(|Jdxd,
In case the equation for
is
given as
F{x,y,z)
ran
The
0,
(21)
{21)
dx dy
(22)
In the above we have assumed that S is such that any line parallel to the z axis
S in only one point. In case S is not of this type, we can usually subdivide S
into surfaces Si, &,
which are of this type. Then the surface integral over S is defined as the sum of the surface integrals over Si, Si,
intersects
The
DIVERGENCE THEOREM
Let
drawn normal
also be written
jjj
i^+^ + lfy^
^j Ardydz
+ A2dzdx + A.dxdy
{2A)
200
A =
Aii
A2J
+ Ask and n =
jrjvAdF
cosi
cos^j
[CHAP. 10
these can
cos yk,
Jj An dS
(25)
STOKES'
THEOREM
non-intersecting curve C
Let S be an open, two-sided surface bounded by a closed
if it is on one side
positive
as
to
S
normal
line
(simple closed curve). Consider a directed
side is positive is
which
of
choice
The
of
S.
side
of S, and negative if it is on the other
sense of C posior
direction
the
Call
advance.
in
arbitrary but should be decided upon
the direction
pointing
head
his
with
of
S
boundary
tive if an observer, walking on the
single-valued,
are
A,,A2,A,
if
Then
left.
his
on
surface
the
of the positive normal, has
t>,
have continuous first partial derivatives in a region of space including
continuous, and
we have
+
In vector form with
simply expressed as
A = Ad +
A23
dA
By
dz
aAi
dAi
dz
dx
(26)
cos a
aA2
dAi
^
|cos^+i-9^--g^ COSy dS
,
+ Ask and n =
=
fA-dr
dAs
cos a
cos
/8 j
this
cos y k,
(VXA)-ndS
JJ
is
(27)
Solved Problems
LINE INTEGRALS
,(1,2)
I.
Evaluate
-y)dx +
(x^
(y^
+ x) dy
along
(a)
(0, 1)
to
(1, 2),
^(0,1)
(6)
straight lines
= t,
(a)
t^
from
(0,1) to (1,1)
(1,1) to (1,2),
(c)
the parabola
+ 1.
An
(0, 1)
and
(1, 2)
in the
xy plane
+ 1. Then
/"*
is
{x^
-(x + 1)} dx +
{(X
+ If +
x}dx
(2x'
2x) dx
5/3
dy
= dx
and
CHAP.
(6)
10]
line
from
f=
^!(
from
line
Since
at
f^_^it'-(t'
2.
If
+ l)dy =
= l,
t=l
+ l)}dt +
at
{(t'
-2/3
10/3
(a)
x=
(b)
(c)
= t^
J^
dr
+ l)dy =
10/3
(1, 2),
+ iy+t}2tdt =
points
to
(0, 0, 0)
(3a;^
(0, 0, 0)
{{3x'
J^
x^t,y = t\z = t\
J^A-dr
and the
j'\2t' + U' +
2t'
+ 2t~l)dt =
)A-dr
evaluate
f rom
(0, 0, 0)
"
to
= t^.
=
If
-2/3
8/3.
t,
dx
{y^
A =
(1, 1, 1)
(a)
(1,1) to (1,2),
{y'
r (x'-l)dx =
+ a;)(0)
(l
and the
Jt
and
(0, 1)
(l-2/)(0)
= l,dy =
(1, 1),
ix'-l)dx
to
(0, 1)
201
and
- eyz)i +
and
and then
(0, 1, 1),
to
(1, 1, 1).
(1, 1, 1).
+ 3xz)j +
{2y
- 6yz) dx +
(0, 0, 0)
then to
(0, 0, 1),
+ Zxz) dy +
(2y
- 6(n{t^)} dt +
(Se
- 6t=) dt +
(4t
{2t'
{1
+ smt')} d{t') +
dt
6t=)
{St'
{dxi
+ dy} + dzk)
~ 4xyz') dz
correspond to
(1, 1, 1)
iSt'
- 4xyz^)k}
(1
and
{1
=1
respectively.
Then
- 4{t){t^)(tY} d{t')
=
12t") dt
Another method:
C, A = (3t' - 6t')i +
+ 2j + 3tk) dt. Then
Along
dr
(i
J A
(6)
dr
(St"
(2t'
+ 3t*)i +
- 6f=) dt +
(1
(4t'
- 4t')k
+ 6t') dt +
J
from
{2(0)
+ 3(0)(^)}0 +
{1
- 6(2/)(l)}0 +
{22/
+ 3(0)(1)} dy +
{1
dx
- 6(1)(1)} da; +
{2(1)
+ 3a;(l)}0 +
{1
A'dr
1-5
0, z
-3.
+ yj + zk =
= 0,dy =
= 1,
da;
(i
f
= 0,
ti+ tH
+ t'k,
dz
while y varies
is
- 4a;(l)(l)}0 =
xi
- 12") dt
- 4(0)(a/)(l)no =
'='
Adding,
- MO)(0)(z')} dz ^
(Bt'
= 0,
from
- 6(0)(z)}0 +
and
= 1,
dy
f'
Jx=o
(
= 0,
2y dy
(3a;^
dz
while x varies
- 6) da; = -5
202
The straight
Then
(c)
is
(1, 1, 1)
= t,
t,
t.
XA
3.
and
[CHAP. 10
dr
{Bf
- 6t^) dt +
+ 31")
{2t
dt
(1
- U*) dt =
6/5
Find the work done in moving a particle once around an ellipse C in the xy plane,
the ellipse has center at the origin with semi-major and semi-minor axes 4 and 3
respectively, as indicated in Fig. 10-3, and if the force field is given by
if
F
=
In the plane z
dr
dxi
(fr-dr
dyj
(3a;
0,
F =
so that the
- 4i/ + 2z)i +
(3x
- 4y)i +
work done
{{3x-'iy)i+ i4x
- 4y) dx +
(B (3a;
+ 2y)i -
[Ax
'i
sin
where
from
varies
to
- Ay'^ + z^)ii.
and
4y'k}
{3(4 cos
x:
*)
a;
4 cos
cos
t)
t}
dt
dt
{4(4 cos
(48t
t)
15 sin* t)\f
OBtt
Evaluate
Since
I
-'c
ds
i/ds
y/dx^
yds
+ dy^ = Vl +
2\/x
Vl +
(y'V
given by
dx
= yl +
=
l/ dx
i/
2\/ic
GREEN'S THEOREM
in the
^/x+l dx
PLANE
from
=3
cc
to x
= ^
(x
):
'Li^E
have
Fig. 10-4
dx dy
- r
P{x,v)\llyl,^,dx
P(x,Yi)dx
156
^^^^'"^^*\
dy
1)""
XX
= 24.
we have
l/xdx,
5.
t j
C we
In traversing
4.
3 sin
Fig. 10-3
t,
Then the
2tt
30 sin
t i
2y) dy
(48
4 cos
+ dyj)
(dxi
+ yi
xi
4y^k
{2xz
is
+ 2y)j-
(Ax
+ 2y-Bz^)j +
{ix
[P(x, y^)
P(x,Yi)dx
P(X, F,)] dx
-JPd
CHAP.
10]
Then
^ p dx
{1)
= -
EAF
JJ ^ dx dy
%
EBF
and
203
be a
and
Xi{y)
Xi{y) respectively.
Then
r rP"
j/S'^''''^
Q{Xuy)dy +
J"
Then
Adding
(1)
and
^ P dx
(2),
J"
^Qdy =
(2)
+ Q
^~&^A^^
dy
( [Q(X.,y)-Q(Xuy)]dy
^ Q dy
Q(X^,y)dy
CC^dxdy
(((y'^-
dx dy.
6.
{2xy-x^)dx +
{x
+ y^)dy
is
of the region
x^,
r
Along
2/^
- x'} dx +
{x
Fig. 10-5
{xY}
(2x^
+ x^ + 2x') dx =
7/6
{2{y'){y)
- {y^} d(y') +
line integral
7/6
{y'
+ y^}dy =
17/15
{y
- 2xy)[f^^^
dx
^1 =
(4y'
- 2y' + 2y') dy =
-17/15
1/30.
\\ {X-2x)dxdy
7.
d(x')
f
=
(a;'
(1
- 2a;) dy dx
-^o
is verified.
Fig. 10-6
1/30
204
STVS
Adding the
hand
left
STVS
ST
and
sides of {!)
SVTS
dx
SVT
TVS
+Q
dy in each
TVSYT
SVT
TVS
TS
dx
TS
we
(2),
ST
Then
%2
SVTS
%i
[CHAP. 10
+ Q dy
(2),
"*"
%,
l)(-j^
J-
)'^^ '^V
II
'^f^
is
proved.
TtJSrT
A region "5^ such as considered here and in Problem 5, for which any closed curve lying in
can be continuously shrunk to a point without leaving %, is called a simply-connected region. A
region which is not simply-connected is called trndtiply-connected. We have shown here that Green's
theorem in the plane applies to simply-connected regions bounded by closed curves. In Problem 10
the theorem is extended to multiply-connected regions.
For more complicated simply-connected regions
as ST, to establish the theorem.
8.
Show
may
it
\ j xdy
given by
is
such
-ydx.
P = y, Q = x. Then
i)
lines,
xdy ydx
T* (^'^
""
a"
jdxdy
(~2/)
dx dy
= 2A
%
where
9.
is
ii
Thus
ellipse
x dy
A =
y dx
X dy
^i
a cos
6,
-|
sin^ e)
de
y dx.
b sin
6.
(6 sin
e)( a sin
e)
de
^277
J2Tr
o-6(cos^ e
abde
vah
*^o
10.
Show
as
shown
in Fig.
is also
such
10-7.
O
Fig. 10-7
AD,
ADEFGDALKJHA
is
simply-connected, and
CHAP.
10]
ADEFGDALKJHA
^'""^
of
if
OEFGD
Thus
J AD ^ "J^;
^^
9^
left,
205
is
the curve
DA
Ci
equal to
is
ALKJHA
ALB:JHA,
Ci, is
the curve
DEFGD
ALKJHA
DSFGD
and C
f + f
is
the boundary
and so
Jt
INDEPENDENCE
11.
of the PATH
Let P{x, y) and Q{x, y) be continuous and have continuous first
partial derivatives at
each point of a simply connected region %. Prove that
a necessary and sufficient
pdx
condition that
dP/dy
dQ/dx
where
is
identically in
Suppose 3P/dy
Sufficiency.
+ Qdy
in
is
dQ/dx
that
%.
dQ/dx.
'^
Necessity.
^^Pdx + Qdy =
Suppose
By
tammg
Green
in
"3^
and that
dP/dy
at
dQ/dx >
at the point (xo,yo).
hypothesis dP/dy and dQ/dx are continuous in
^, so that there must be some region r con{x,yo) as an interior point for which dP/dy - dQ/dx >
0.
If r is the boundary of r, then by
'
dP/dy
theorem
cannot be positive.
we can show
Similarly
identically zero,
12.
Let
and
dition that
is
that
i.e.
dQ/dx
If 3P/dy
for
aZi closed
curves in
yPdx + Qdy
=
that
dP/dy
Sufficiency.
dP/dy
P dx + Q dy
^
Thus dQ/dx
-
follows that
identically in
it
??.
^
Pdx + Qdy
A-
ADBEA
(see Pig.
From
10-8).
Pdx + Q dy, we
this,
have
/^/"'
"">'
i.e.
the integral
is
"'^
ABB
= ~ f " f
BBA
AEB
and so
r=r^
^"l
"^
must be
sufficient con-
joining points
%.
it
- dP/dv
/"
and
206
[CHAP. 10
Necessity.
and C2
X
From
X'
13.
and
(a)
be as in Problem
differential of a function
(o)
is
zero,
ADBEA
in
and hence by
11.
Pdx + Qdy
sufficient condition
(j>{x,
dP/dy
y) is that
Pdx + Qdy
we have
AEB
that
dQ/dx.
d^
^A
I
and
dQ/dx.
integral around
Let
and
in
,j>{B)
be an exact
where
^(A)
Necessity.
Pdx + Qdy = d^ =
If
an exact
and
(2)
^dy,
dy
dif>/dx
(1)
(x,y).
P,
*{x
+ Ax,
y)
B,/,/dy
(2)
then
differential,
^dx
+
dx
con-
Sufficiency.
By
\
Prob.
12,
P dx + Q dy
is
dP/dy
if
dQ/dx,
then
(a.h)*
and define
Fig. 10-9
P dx + Qdy
(a, b)
Then
Pdx + Qdy
(a, b)
Pdx + Q(
Pdx + Qdy
Since the last integral is independent of the path joining (x,y) and (x + &.x,y), we can choose the
path to be a straight line joining these points (see Fig. 10-9) so that dy = 0. Then by the mean
value theorem for integrals,
.f.(x
+ Ax,y) -
Thus
(6)
Let
it
^f^ + Ajj/)
-* 0,
we have
we can show
B=
that
P dx ^
d<(>/dx
Q.
d(f,/dy
(xa, 2/2).
From part
^ (.x^,y{>
P dx
-\r
0<ff
<1
P.
Pdx
-f
dif,.
Qdy
Q dy, we have
<p{xi,y2)
'^ (.a.bl
eAx,y)
(a).
<f>{x,y)
^A
P{x
P dx + Q dy = -^dx + ~dy =
follows that
{xi,yi),
= J_ f
.f>{x,y)
^ ia.hl
<p(xi,yi)
<I>(B)
^(A)
CHAP.
10]
207
/-C3,4)
14. (a)
Prove that
{Qxy^
dx
y^)
{Qx^'y
Zxy^)
dy
is
(1.2)
P = 6xy^-y\ Q =
line integral is
(6)
Method
(6)
- 3xy\
Qx'y
Then
dP/dy
12xy
(a).
- 3y^ =
dQ/dx
Since the line integral is independent of the path, choose any path joining
(1,2) and (3,4),
for example that consisting of lines from (1,
2) to (3, 2) [along which y = 2, dj/ = 0] and then
(3,2) to (3,4) [along which x = 3, dx = Q]. Then the required integral equals
-S)dx +
(24a;
J
Method
(54^/
- 9y^) dy =
(6xy^-y')dx
(6x^y
=
f(y)
- Sxy"-) dy =
236
Note that
We
- y^) dx +
{x^ycosx
x^^^
cos X
+
+
y^'^
(6a;V
clear that
it is
^
x^'y
way
in
Hence
- a;^/' + c)
(i(3a;V
+ c|";^J =
can be omitted.
236
16,
Page
115.
from which
hypocycloid
only
g{x)
3a;V-x2/'
{6xy'
Evaluate
3x'y'
P =
156
2:
15.
80
+
-
{Gxy" dx
rf(3y)
6x'y dy)
d(xy')
(y^
dx
d(3x^y^
3xy^ dy)
xy")
= 3xV xy^ + c.
<f>
+ 2xy
=
=
3xy^) dy
sinx
y^e-)dx
{x^sinx
2ye^)dy
around the
a^'^.
2xy sin x
Q ~
y'^e",
a;^
Then
bPlby = X' cosx + 2x sin a; - 2ye''
around any closed path, in particular x^'^ + y'"^
sin
=
=
a;
2ye''.
dQ/Bx,
a"^,
so that
by Problem 11 the
line integral
zero.
is
SURFACE INTEGRALS
16. If y is
is
VF-k =
from which
|secy|
VTTWT4
|sec,|
F(x,y,z)
is
^"^
~^"''
Fy
'
equation
Zy,
F^
=l
is
or F{x, y,
f{x, y)
a normal to
|VF||k|cosy
F.
or
at {x,y,z)
= V^l +
z)
F|
^^
0.
VF = FA + Fyj+F.y..
is
the
Then
Fi cos y
as required.
|F.|
^In casejhe
0,
= y^'+Jf +
S and
and we
f{x,y),
find
we can write
[see y|
\/l
zl
F(x ,y,z)
z}.
z-f(x,y)
0,
from which
208
17.
ji
Evaluate
U{x, y,
dS where S
z)
integral
11 U{x, y, z)
where
(6) x^
(a) 1,
+ y^,
= 2-{x^ + y^)
Give a physical
(c) 3z.
equal to
is
\/l
is
equal to
is
[CHAP. 10
zl
zl
dx dy
(1)
%
S on
the projection of
is
given by x^
Since Zx
\(
y^
0.
z, = 2y,
2,
2x,
U(x, y,
z)
the xy plane
Vl +
(i)
4.T=
can be written
+ 4^ dx dy
(2)
%
(a)
If ll(x, y, z)
1, (2)
becomes
+ 4x^ + 4y^ dx
To evaluate this, transform to polar coordinates (p, 0). Then the integral becomes
/2ir
V2
I
^(l + 4pV
VTTVpdpd0 =
X2T-
(6)
U(x,y,z)
If
x^
+ y^
(2)
becomes
jj
+ y^Wl +
4a;''
p'VT+Vdpd0 =
{x'
to p is accomplished
4)/^
S assuming
unit density.
or in polar coordinates
dx dy
149)7
30
by the substitution
S about
Vl + 4p' -
u.
density, or the
(c)
U{x, y,
If
z)
f(
3z, (2)
3Vl
becomes
4x^
4y' dx dy
Jj
3{2
(x'
+ y')} Vl +
4x^
4y' dx dy
density
32, or
or in polar coordinates.
3p(2
111,7
(
10
S assuming a
moment
first
18,
- p^) Vl + 4p' dp
ax).
Since
Zx
=
Va^
we have
x^ y^
and
2s
=
V^^^
x''
1/''
Fig. 10-11
CHAP.
10]
Two methods
Method
^/l
JJ
x''
7r/2
(f>
The
2:
is
a cos
p dp dtp
dx dy
4>,
v^^
integral
tan'' e,
d(f>
1^
p=
2a Xrr/2
(1
sin
(tt
<p) d<t>
- 2)a^
equal to
is
^^9 =
Va'
=r:r dy
a;'
da;
2a
x:
^=o
2/^
this integral
becomes
tan
de
costt>
2a
(9
sec^ e
=
a;
)
^p
*^a: =
Letting
dx dy
z^
/%acos<l
=
Method
zl
Since
2a
4an J^
2a^|fltan^eC'*
2a^{:7/4
-/aa;
sm"
sin
da;
os'
Va +
da;
a;
tan^ $[
(tan 6
(sec^
- e)C^
1)
(^
de I
Note that the above integrals are actually improper and should be treated
limiting procedures (see Problem 78, Chapter 5, and also Chapter
12).
19.
dS
{^'^^
^Jl/M.
SS
zVTT4^+4y^ dx dy
J J A-ndS,
Evaluate
plane 2x
A
and
+ 2y + z =
normal
to
is
V(2a;
where
A =
xyi
included in the
2y
^ All
130
137r/3
20.
+ z-6) =
2i
x^j
first octant,
(x
+ z)k, S
and n
is
is
a unit normal to S.
+ 2j + k,
^ ^i2ik_^ 2i + 2i + k
3
V2' + 2^ + 1^
A-n = {xyi - a;'j + (a; + z)k} ^ Ji + ^j + k
_ 2xy - 2a;' + (x + z)
so
2xy
2x'
(a;
+ 6 - 2a; - 2y)
2xy
2x^-x-2y + Q
3
is
by appropriate
_ %
" SSVl + ix^ + iy'dxdy
2)a^
Problem 17
in
Sf z
tively,
209
therefore
X'
Fig. 10-12
210
jj
^xy -2x^-^x-2y + 6
(2x2/
I
x=o *^y =
+ 6y)|^
(a;2/''-2x^i/-X2/-2/''
two-sided.
^^
"'dx
27/4
f.
AD
The
22.
^^^
- 2x* - X - 2j/ + 6) % dx
J.1 =
21. In dealing
^,
^,
^,
^^ ^ ^
[CHAP. 10
Fig. 10-13
DIVERGENCE THEOREM
'>
11)
Fig. 10-14
Let S be a closed surface which is such that any line parallel to the coordinate axes cuts S in at
most two points. Assume the equations of the lower and upper portions. Si and S2, to be z = fi(x,y)
and z = f2 (x, y) respectively. Denote the projection of the surface on the xy plane by "5^. Consider
8 As
dz
dy dx
dz
%
JJ
A3{x,y,z)
dy dx
jj
[A3(x. y,fi)
As(x,y,fi)]
dydx
51
S2,
dy dx
cos y^ dSi
nz to S2
makes an acute
CHAP.
10]
Then
- -cosy^dSi = -k-nidSi
dy dx
Si,
Ij A3(x,y,f2)dydx
A3(x,y,fi)dydx
makes
nz dS2
S2
'
JJ
A3 k
211
1) A3 k
ni dSi
and
A4x,y,f2)dydx
j^i^
JJ
A^ix.y,
dy dx
ft)
A3 k
JS2J
n2 dSz
+ j
A3 k
dS,
'sj
jj Ask-ndS
so that
(^)
Similarly, by projecting
Adding
and
(1), (2)
XfA3k.dS
fff't'^^
{3),
ffK^-'W^ty
or
V A dy
JjV J
//(Aii + A2i+A3k)..S
=
'
I'l'
n dS
The theorem can be extended to surfaces which are such that lines parallel to the coordinate axes
meet them in more than two points. To establish this extension, subdivide the region bounded by
S
into subregions whose surfaces do satisfy this condition. The
procedure is analogous to that used in
Green's theorem for the plane.
23.
We
first
evaluate
Face DEFG:
JJ
A-n
dS
i,
in
xr^- dS
=
z
where S
is
the
dj/
xz^k
Fig. 10-15.
Then
x=^l.
- 2)i +
= J J
"
{(2
{2-z)dydz
z^k}
. i
'^
DEFG
Face ABCO: n
JJ
-i, x
n dS
=
=
0.
3/2
Then
j'0 j
^0
^0
^0
i-zi)
'
zdydz
(-i)
dy dz
1/2
Fig. 10-15
212
Face
ABEF:
= j,
= 1. Then
2/
llA-ndS=
Face OGDC:
n=-i,
{{2x
- z)i +
A-ndS =
\\
= k,
x'-j
{(2a;
dx dz
dx dz
x''
1/3
{(2a;
z)i
X2^k}
(-j)
dxdz
- l)i +
a;Vj
- xk}-kdxdy =
J
a;
da; d^/
-1/2
Then
0.
A-ndS
l\
Then
1.
jJA-ndS = J J
n = -k,
xz'Vi)
"
BCDE
Face AFGO:
Then
0.
OGDC
Face BCDE:
[CHAP. 10
- x^yj}
{2x\
(-k)
da; dj/
AFGO
Adding,
Ci"
A-ndS = f+^ +
O-l +
^.
Since
\\\ V A dy
24.
Evaluate
{2
+ x^- 2xz) dx dy dz
is
dS,
where S
is
a closed surface,
By
jjr.ndS
jjf^-^dV
V
where
25.
is
Evaluate
xz^
dy dz
(x^y
z^) dz dx +
{2xy
+ y^z) dx dy
where S
is
the entire
Since
(a)
dS
cos
a,
{xz^ cos a
dz dx
{x^y
= dS
cos
/?,
z^) cos f3 +
da;
(2xy
dy
= dS
dS
where
A =
xz^i
(x'^y
z^)i +
(2xy
y^zyk
and
cos a
cos
/3 j
is
(a)
by the
n dS
cos y k,
where
cos y,
y^z) cos 7}
y^ and
the outward
CHAP.
10]
By
XTT/2
^Tr/2
9, this
integral
213
equal to
is
y^a
r'
r^ sin
dr de
d<f,
5
If Si is the
(6)
{X'y
JJ
ffxz^dydz
- Z>) dzdx
f^'^zWa^-y'-z'dzdy (
'"''
{X'y/d'-x^-z^
+ j/^^)da;rf2/
(2a:2/
xz'dydz
JJ
jj(2a.2/
f"'
+ 2,^.)da;d2/
jj
*2
(a;'2/
- =) dz da; =
{2a;2/
+ 2,^(0)} dx d,/
(T
0,
{20=2/
3/
"^
we
j/Va^
^3}
a,^
{z
f''^
d^
= 0),
then
.^^a"
- y^ ~ z' dzdy
rf^
a;^
- 2^ -
dz
da;
- j,^} ^^ ^^
0,
C^"^^
"^ x=-a
Sa
By
{-arVa'
=-a ^2 =
JJ
the base
is
2xy dy dx
^3=-Vo2-i2
obtain
z^
dz dx
.Va2-i2
-x'~y' dy dx
Since by symmetry
2/Va^-a;^-y^d2/da;
12
f" C
p'
is,
^W
4,
V^F^ p dp
d<i>
^^
THEOREM
STOKES'
26.
all
Let
Assume S
10-16.
to have representation
or y-h{x,z), where f,g,h
are single-valued, continuous and differentiable
functions. We must show that
z
= f(x,y)
JJ
or x
= g{y,z)
(VX A)'ndS
=
Jj
s
j
[V X (Aii
+ A^j+Aak)] -ndS
A'dr
dx dy
where
is
the boundary of S.
Consider
first
^^
[V X
(A,i)]
n dS.
Fig. 10-16
214
VX
Since
Ml
A i_
A
dx dy
(Aii)
[CHAP. 10
aAi,
ay
dz
dz
Ai
If
is
f(x, y)
yj+zk =
xi
+ yj+
f{x,y)k
dy
so that
= ^'^^^ ^ ^^%^'
|^
^"*
is
a vector tangent to
is
n-j
T n-k
= dz
r n-k
dy
,
n-j
3y
to obtain
'3Ai
~ n-k ]dS
dy
dz
Now
a; i
to n, so that
(1)
(i)
Substitute in
-Mtn-k^s
= (^Mln.j
[VX(A.i)].ndS
_
^
[VX(Aii)]-ndS
on S, A,(x,y,z)
[V X
A^lx,y,f(x,y)]
-n dS
(Aii)]
Ml n'k) dS
BAi dz
^
dz By
/Ml
dy
Ml
dz^\
()
\Sy
F(x,y);
n-kdS
= -
^+^^ =
hence
|^ and
(2)
becomes
= - - dx dy
Then
rr
[V X
(Aii)]
n dS
where
is
'?i
the projection of
^ F dx
equals
where r
is
-j-dxdy
\\
S on
By
the xy plane.
the boundary of %.
is
is
the
we
{x, y)
of
r the value of
or
[V X
jj
(All)]
-ndS =
j)
Aidx
rr
[V X (Ao)]
n dS
A^dy,
Thus by
[V X (Ask)]
J* J
n dS
A, dz
addition,
rr
(V X A) n dS
9 A
dr
The theorem is also valid for surfaces S which may not satisfy the restrictions imposed above.
.,Ck which
Sk with boundaries Ci, d,
For assume that S can be subdivided into surfaces Si, S2,
do satisfy the restrictions. Then Stokes' theorem holds for each such surface. Adding these surface
integrals, the total surface integral over S is obtained. Adding the corresponding line integrals over
.
Ci, Cz,
is
obtained.
. ,
CHAP.
10]
27.
is
215
dr
J)
Zydx - xzdy +
yz' dz
t)
dt
(2
dt
t)
'277
VXA =
Also,
+ 8cosH)dt =
(12sin^
_fl_
_a_
_a_
dx
dy
dz
207r
{z'
x)i
(z
+ 3)k
Fig. 10-17
3y xz yz^
and
xi + yj k
\/^+ y' + l
+ y'-2z)
+ 2/^-2z)|
V{x'
I
V(a;^
Then
JJ(VXA).ndS =
=
JJ(VXA).^
ffH^y
+tVl
x'
fjixz^ +
x^
+ z + S)dxdy
+ sUxdy
28.
Up
is
Sufficiency.
that
Suppose
A=
VXA =
cos 0)(pV2)
p2 cos^
+ p72 +
j>
<
d0
2077
A-dr =
*^c
identically.
0.
3} p dp
A'dr
(VXA)-ndS =
JJ
s
Necessity.
Suppose
_^^A.dr
VXA #
at some point P.
Then assuming: VX A is continuous there will be
a region with P as an interior point, where
V X A ^ 0. Let S be a surface contained in this region whose
normal n at each point has the same
direction as V X A i.e. V X A = an where
is a positive constant. Let C be the boundary of S.
Ihen by Stokes' theorem
that
VXA=
is
^A
dr
also a necessary
and
VXA=
sufficient
>
0.
condition
for
a line integral
216
29.
Sufficiency.
A=
If
VXA=VxV0 = O
V^, then
Necessity.
is
by Prob.
80,
A =
Chap.
7,
is
Page
that
A=
we
VXA =
0,
(B
v^-
158.
dr
If
[CHAP. 10
A'dr
and
(o, b, c)
(x, y, z).
and
dt
Let us define
Then
+ ^x,y,z) -
<t,(x
<t>{x,y,z)
Aidx + Aidy + At
dz
Since the last integral is independent of the path joining (x,y,z) and (a: + Aa;, j/,z),
the path to be a straight line joining these points so that dy and dz are zero. Then
(a:
+ Aa!.
3/.
we can show
Similarly
Thus
that
d(p/dy
A = Aa + A.j + A.k =
30. (a)
(&)
Show
Necessity.
integrals.
d<f>/dx
B<l>ldz
Ai.
As.
gi+gj+lfk =
V^.
yxA
A =
where
If
Aidy + Aidz =
-<r
IS = ^^
Then by
differentiating
we
dAi
By
is
Aii
+ Aaj + Ask.
from which
Sufficiency.
If
_
~
aA%
dx
(^)
d0
eAi_
'
VXA =
0,
'^^'^
^^
^{x^.y^.z^)
"*"
af''^
(^)
VXA =
dAi
dAz
dy
dz
~^^''
<i,{xi,yuZi)
*^"
^'
fz =
'
_ Ms
dx
dz
0.
VXA=VxV0
If
Another method:
d<^
which
Ai,
< < 1
+ eAx,y,z)
Ai{x
Aidx
(a)
Aidx
Aa;^0 gives
sides as
d^,
then
O.
A=
V<f>
and
choose
J,
we can
^^^2/
+ ^d =
d0
d^,
CHAP.
10]
31. (a)
F =
Prove that
field.
Evaluate
(6)
(2xz^
+ 6y)i +
(Sx
F- dr where
J^
+ A^dy +
A,dz, we have
~ 2yz)i +
(Bx^z^
- y^^i,
is
217
(1,
i,
-1,
a conservative force
1) to (2, 1,
-1).
(c)
Give
force field
any two
is
conservative
points.
VXF =
Since here
dy
dz
Method
From
+ 6y
6a;
be conservative
- 2j/z
we
x^z^
Gxy
1,
/. (y, z)
F-dr
we may
Alternatively
F-dr
from which
0,
x'z^
"
is
conservative.
3a;V-j/^
6x2/
?/^.
is
=
=
+ h (x, z)
- r^^ Az\ ~
t (^
'
..\
^^^"'^^
y^z
f, (x, y)
"
'''''"'
^^''^'^
^"^
x'z'
6xy
y'z
c\';-'\ll
15
(2xz'dx
d(x^z')
+ 3xVdz) + (&vdx +
+ d(6x2/) - d(y^z) =
6xdy)
d(a;2z3
- (2yzdv + y^dz)
+ g^^ _ ^2^ ^ ^^
determined.
2:
*^
r,rt,VnJ"'^
'"*T*^
'^
J^^^(2x-6)dx +
J^^^^
(12-22/)d2/
we can
\l2z^-l)dz
15
^^ ~ n j
Y
^ " 2, .r-i
1, dx - 0, d. = 0; and the third integral
=
=
x
2,y X. dx =
=
x~2
0,
Physically
J^F-dr
dy
work done
is
(1,
-1,
Q dz
dz~a- U,
- 0;
by placing
1) to
(2, 1,
-1)
joining these
MISCELLANEOUS PROBLEMS
^^'
0.
-O
X(2,
along C.
joining
obtain respectively
(c)
VXF =
that
is
1:
these
Method
is
dx
2a;z
(6)
jj>dr
if
"" "
^." ^^"'""^ "^^^"^^ ^ transformation which maps a region g? of
?i
i^"^'"^'
the
^
xy if
plane mto
a region %' of the z^t; plane, prove
that
218
J/|^ dudv
JJdxdy =
(b)
(a)
(a).
is
-^xdy
dxdy
[CHAP. 10
8,
ydx
%
Under the given transformation the integral on the right of
liO
S*)-'(*+S-)
is the mapping of
simple closed curve also).
where
By
Green's theorem
if
is
<>
IX.(^-'S>"-(l^-)*
in the
%'
becomes
(1)
to be such that
C, the right
is
side of (^)
equals
dx
(x^
dv\ du
ff
Idx
dx dy\ J
dy
=SJ
CClH^jyX du dv
J J \d{u,v)
dxdy.
%
we can show
In general,
Problem
(see
jjdxdy
and
%'
du dv
J/ ||g^
(S)
%.
(6)
|M dudv
jj F(x, y) dx dy
83) that
the
first
expressed in rec-
33,
Let
F =
T'^V
+
yxF.
Calculate
(a)
i F-dr
Evaluate
(b)
-'
(o)
_a_
VXF
dy
dx
in
(0,0).
dz
-y
x^
(6)
C^F'dr =
+ y^
x'
+ y^
ydx + xdy^
.^^^
pcos4>,
j/
p sin 0,
where
(p,<p)
Then
dx
p sin
d0
y dx +
and so
x'
= 27r
ABCDA
dp cos 0,
X(
+ y'
dy
p cos
d(.arc tan
(f>
d<f,
dp sin
</>
=
J
at
d<f>
A
=
and
Ztt.
CHAP.
10]
219
(a)
(b)
Fig. 10-18
- 00
Q =
F = Pi +
VXF =
Qj,
to contradict those of
Problem
is
equivalent to
dP/dy
However, no contradiction
11.
x'
+ y^
exists since
P =
Problem
,i
= ^
<
at
0.
BQ/dx
was assumed
34. If div
Since
seem
PQRSP
-~
+
and
and
this
x^
(0, 0),
y'
11.
field
at a point P,
show that
SSA'tidS
A =
div
lim
AF
AV-tO
aV is the volume
aF to the point P.
vi^here
ing
By
div
AS and
A dV =
AV
By
the limit
is
obtained by shrink-
n dS
i
AS
the mean-value theorem for integrals, the left side can be written
A iijdV =
div
where div
Then
is
div
A AV
AF^O
such that
is
A =
always interior
SS
A =
throughout AV.
AV
to
point P; hence
div
n dS
lim
Av-^o
at
A-ndS
AV
This result can be taken as a starting point for defining the divergence
of A, and from it all the
properties may be derived including proof of the divergence
theorem. We can also use this to extend
the concept of divergence to coordinate systems other than
rectangular (see Page 142).
Physically,
ijjj A-ndSj/AV
AS
represents the flux or net outflow per unit volume of the vector
A from the surface AS. If div A is positive in the neighborhood of a point P it means that the
outflow
from P IS positive and we call P a source. Similarly, if div A
is negative in the neighborhood of P
the outflow IS really an inflow and P is called a sink. If
in a region there are no sources or sinks
then div A =
and we call A a solenoidal vector field.
220
[CHAP. 10
Supplementary Problems
LINE INTEGRALS
35.
Evaluate
from
lines
Ans.
36.
f'"
(x
to
(1, 1)
+ y)dx + (y- x) dy
along
and then to
(d)
(1, 2)
(6) 11,
(a) 34/3,
(c)
^ (2x - y +
Evaluate
4)
dx
the parabola y^
the curve
2t^
+ 3a; - 6) dy
(5y
38.
(a)
y =
Ans.
39.
F =
(x'
- y')i +
x'-x from
2xyi,
the point
evaluate
(1, 0)
a straight
(b)
1,
t'
+ 1-
to
J"
the xy
in
F =
(2x
where C
y) ds,
is
x'
+ y' ^
41.
If
(a)
the
(a)
23/15,
s is
the
Ans. 15
is a path
+ 2z)i - x^k. evaluate J^F-dr from (0,0,0) to (1,1,1), where C
points, (c)_the straight
curve x = t,y = t\z = t\ (h)\ straight line joining these
and then
to (1,1,1),
(d)
x^z,z^y.
the curve
13/30
{d)
is
and
25
(y
from
Ans.
(3x-22/)i
consisting of
lines
plane given by
124/15
(a)
Evaluate
If
(0,0).
(6)
(2, 2).
with center at
circle of radius 4
dr
in the
40.
Ans. 12
37.
If
+t+
around a triangle
= x,
32/3
(d)
14,
(4, 2),
(a)
J^
dr
F -T
where
ds
is
a given force
field,
prove
is
J^
GREEN'S THEOREM
42.
in the
PLANE. INDEPENDENCE
43.
Evaluate the
44.
(a)
Let
line integrals of
Problem 36 and
/.
d) (o-ix
+ a^y + da) dx +
p cos 0,
P sin ^,
prove that
46;.
If
47 .
J/
48.
,,
(a)
Prove that
(2,1).
{y'
{h)
x^
y^
(a).
h^y
+
,
63)
{x^
j
dy
x^
2/^
- 4xy^) dy
Ans.
(h)
=
is
=
C
Stg
2/
x dy - y dx
and
is
a square with
y-x^y-)dx +
where C
-2xy) dy
(6)
(b^x
- xy^) dx +
Prove that
having area A.
be any simple closed curve bounding a region
are constants,
(6)
(a)
PATH
45.
^^{x^
of the
:L
xy^ dy,
16.
ti.
(61
\a
- <h)A
Ans.
be zero?
(b) a,
b.
,,
2^.]
,^
A^
Ans. Z^a /8
and interpret.
where C
is
120.
(1, 0)
and
CHAP.
10]
Evaluate
49.
to
(0, 0)
dx
(1
= Ty'
from
(0,0),
2y sin x
+ SxY) dy
2x
J-V4
Ans.
51.
y' cos x)
Atis.
(ff/2, 1).
Evaluate the
50.
(2xy'
J^
221
SURFACE INTEGRALS
52.
Evaluate
(a)
{x'
^j
+ y^)dS,
where S
is
z"
3(a;"
+ j/")
bounded by z
and
(b)
3.
Ans.
(a).
53.
54.
55.
56.
57.
a)
An^.
<
J/",
< :r/2.
(a) 2jra"(l
(a)
Find
a),
t he
<
Ans.
a;"
+ 2/" + z" =
a;"
2/"
a"
< 7r/2.
ia(l
(a)
(6)
From
cos
a),
a;"
the result in
(b) a/2
Va;"
+ 2/" + z" -
a"
= 3,
J/".
x''
and
x'
+ z"
a"
= 2,
2/".
z tan a
(c)
o".
Explain
(a)
DIVERGENCE THEOREM
A =
= 0.
Evaluate
Js J
bounded by
62.
+ j/")
is
by
61.
3(a;"
which
2/"
= \/^M^,
hemisphere.
60.
cos
z tan a
The
a;"
x-0, y = 0,
(a)
Use the
(6)
formally placing a
cut off by
why
59.
z"
Va;"
58.
2z
(a) djr
Evaluate
F-ndS,
2/",
jj A-ndS,
where
a;
= 0,
a;
F =
A =
where
(2xy
(z'
+ z)i +
2/"j
- x)i -
xyj
+ 32/)k taken
= 27
+ 3z)i -
{xz
and S
Szk
(x
is
the
surface of the
region
Ans. 16
+ y)i +
(2/"
+ 2z)k
and S
is
(3,
-1,
2)
and radius
3.
Ans.
lOSn-
where
is
the
(6) directly.
a;"
+ 2/" =
Ans. Slv
and
= 3,
(a)
222
64.
ff 4xz dy dz -
Evaluate
^ = 0, z =
Ans. 3/2
= 1,
0,*a:
65.
Prove that
66.
Prove that
67.
If
2/
= 1,
y^ dz
dx
= 1,
where S
yz dx dy,
(a) directly,
fj (V X A) n dS
is
by Green's theorem
(fe)
[CHAP. 10
a;
= 0,
ndS =
where n
0,
drawn normal
the outward
is
to
drawn normal
is
to
jJJdivndF = S
V
STOKES' THEOREM
68.
a;2
69.
70.
A =
+ z"
j^2
and C
2yi
^ Zx] - z'Xa,
boundary.
is its
Evaluate
JJ
VX
A)
\ =
n dS, where
(x
- z)i +
(x'
+ yz)j -
and S
Sxy^k
is
s'
cone
71.
If
= 2-
is
y/x^
+ y^
Ans.
12it
S and B = V X
A,
jj B'n dS =
prove that
0.
72.
Prove that
(a)
F = V0.
Ans.
73.
x'
74.
Let
- 42/ +
3x
Show
b\
- 4z)j C
4j/k
a conservative force
is
(c)
is
if
f F- dr
F=
if
5r'r,
F=
(h)
field,
to
(2, 1,
a-
to
Find
4>
such that
-1).
where r
/(r)r,
x'
where
+ y' + z^ =
a;i
+ 2/j + zk,
/(r) is
then
assumed
J^
dr
ft'^- o=.
to be continuous.
r f(r) dr
a function
is
<p
such that
F =
F =
Ans.
(x^
+ constant,
that
F-dr, where
In Problem 73 evaluate
(a)
76.
x^j/
3)1
+ y' + z^ =
Ans.
75.
(2xy
Evaluate
(c)
(6)
F =
F = V^,
where:
(xz
(z' - Axy) dx
Solve the differential equation
Ans. xz' - 2x^y + 3y' + z = constant
(6y
find
it.
constant
- 2x') dy +
(Sxz'
+ l)dz =
0.
MISCELLANEOUS PROBLEMS
77.
two, at least)
78.
in a region
is
that
(where
-^ + -^
is
^^^dy-j^dx
order
continuous and has continuous partial derivatives of
0.
10).
CHAP.
79.
If
10]
Pdx + Qdy
223
Ans.
80.
81.
(a)
p.
e'^^"'^'
(6)
= cx^-x,
where
x^
y'
c is
any constant
+ {z- aY =
82.
Prove that
position
83.
VX
Js J
and n
is
(0n)
yjx^
+ y' + z'
x^
+ y\
prove that
dS =
a unit outward
a'
where ^
drawn normal
any continuously
is
of
to a closed surface S.
from sign
^^ F(x,y)dxdy
^ Q(x,y)dy
Q^^fiu.v),
J^
theorem to transform this into
84.
Flfiu.v), g(u,v)]
\\
JJ
F(x,y).
g{u,v)f^du + ^dv'].
^i^^ldudv.
B(u,v)\
a; = f{u, v, w),
y = g(u, v, w), z = h{u, v, w) defines a transformation which maps a region
xyz space into a region %' of uvw space, prove using Stokes' theorem
that
valid.
F(x,y,z)dxdydz
where G{u,
IS
v, w) = F[f{u,
See Problem 83.
G{u,v,w)
jj
d{x, y, z)
d(u, V,
w)
of
du dv dw
ds^
E =
where
86.
dQ/dy
If
JJJ
85.
y,here
^
OW
'^
bu
F=^-^
G=^.^
du
dv
'
dv'
dv
(a)
(b)
Deduce from
(a)
t{u,v)
is
is
given by dS
= y/EG-F'dudv.
fC y/EG-F^dudv.
s
[Hint:
(A X B)
87.
88.
(C
||
X D) = (A C)(B D)
=
(b)
0| =
V(^^1>(S^
(A D)(B
^""^
'""^^
"^
*^^
*''-*-
C).
to obtain
divA
in
(a)
cylindrical
and
(6)
spherical coordinates.
See
chapter
11
Infinite Series
CONVERGENCE
Consider the
and
DIVERGENCE
INFINITE SERIES
of
infinite series
'^ U
Ui
U2
Us
n= l
= Uu
S2
Ml
Sz
M2,
=^
Ui
U2
+ Us,
and S
called its
is
where
.
.
= U1+U2+
Sn
...,
there exists a
i.e. if
+U
limS
S,
(2)
the
sum.
divergent.
Example
Example
1:
2|r =
f+i + i+--
Chap.
Then
2:
3).
(-1)"''
lim
since
Here S = sum
Tl -
n-+oo \
=
^^
^ /
= 1-1 + 1-1+--.
exist
1,
the series
Here S
and the
of first
n terms
is
- 1;
or 1 according as n
is
S=
1.
even or odd.
series is divergent.
n-+oo
If
limM
2 converges, then
ever
is
i.e.
lim Un
0,
Chap.
2m
3).
may
or
may
not converge.
is
divergent.
2.
3.
Removal
number
of terms
from
SPECIAL SERIES
1.
Geometric series
J ^""'
a + ar + ar^+...,
n=l
converges to
terms
is
S=
= '^^^^
if
\r\
(see
< 1 and
diverges
224
if
3).
\r\
^ 1. The sum
of the first
CHAP.
11]
INFINITE SERIES
*
The p
2.
>
?>
TESTS
series
J^
The
^here p
series with
DIVERGENCE
and
Let
O^Un^Vn
t;
for
go
for
is
=1
is called
SERIES
of
of
CONSTANTS
Example:
n>N
N = l.
Often,
^^ g 1
Since
all
n>N,
all
(6)
Convergence.
if
Comparison
(a)
+ __+_+...
CONVERGENCE
for
1.
225
2 |r converges, ^~r^
and
also converges.
Since
>
2^
and
diverges,
Then
also diverges.
2.
(a)
If
^o and
i;gO and
lim
if
J^
A^O
or
=,
If
(c)
If
to
Theorem
it.
1.
A=6
A = 00
in (a)
Let
lim%''M
(i)
lun converges
(w)
2m diverges
Examples:
3.
in (a)
if
if
1.
>
2:^
pg
-^
2i
and
^ converges since
Inn
mn
.
A is finite
A v^ (A may
and
.^
"S,
Then
A.
,.
VV+T
infinite).
lim n'
-^
In
diverges since
lim n}
--
~.
(n+l)'
such that
f{n)^un, n
= N,N+l,N+2,
fix)
dx
^m
as
is
Example:
4.
be
-j converges since
fix)
Hm
dx
converges or diverges.
^=
Hm
is
-^
In particular
exists.
[CHAP.
INFINITE SERIES
226
An
(see
Problem
if
15).
For the
Example:
11
\u+i\
(b)
lim u
ior
\Un\
n^l
(or lim
series
=-,
\un\
(a)
\un+i\
=^.
Then for a
2^
\un+i\
1,
\un\
O)
^
^
|m,|.
we have
Also
lim
Mn
|m|
=0. Hence
Theorem 2.
The numerical error made
we
If
than 1
5,
in
l-i + i-j + i
is less
= 0.2.
Absolute and conditional convergence. The series 2m is called absolutely convergent if 2 \un\ converges. If Swn converges but 2 \un\ diverges, then 2m is
called conditionally convergent.
Theorem
3.
is
Example
1:
T?
2:
Sm converges.
+ 4-|?-J5-+^ + ^--is
+ 3?+45"+
^-^ + ^-^+
is
converges, but 1
converges.
+ 4+
diverges.
Thus
2"
~
3
conditionally convergent.
Any
of the tests used for series with non-negative terms can be used to test
6.
Ratio
Un+l
lim
converges (absolutely)
(6)
diverges
L=1
if
test.
Let
converges (absolutely)
(b)
diverges
if
if
series
2m
L<1
fails.
(a)
L=
Then the
L>1.
the test
If
8.
Let
(a)
If
7.
test.
lim y/juj^
if
Then the
L.
series
2m
L<1
L > 1.
Raabe's
test.
Let
limw(
'^
L>
(a)
converges (absolutely)
(b)
if
^-
Then the
1
if
L<
1.
series
2m
CHAP.
INFINITE SERIES
11]
L=1
If
the test
This test
Gauss'
9.
(6)
If
fails.
often used
when
L c
'^~n'^^'
Un + l
vtn
2m
series
(a)
test.
is
"^^^^^
This test
THEOREMS
on
is
227
if
\Cn\<P
for
all
n>N,
then the
LSI.
Theorem 4.
The terms
Theorem 5.
The sum, difference and product of two
convergent. The operations can be performed
UNIFORM CONVERGENCE
of
absolutely
is
FUNCTIONS.
IS
and each x
\Un{x)-F{x)\ < . for all n>N. In such case we
write hmun{x) = F{x).
The number
may depend on x as well as e. If it depends
only on c and not on x, the sequence is said to converge
to F{x) uniformly in la, b] or to
be uniformly convergent in [a,b].
we can
[a,b]
find
iV>0
such that
The
^U{x)
n1 = l1
is
Ui{x)
U2{x)
Uaix)
...
where S{x) =
(3)/
V
{S(a;)},
we can
find
and not on
iV>0
x,
<
for
all
1,2,3,
In such case
we
write
>0 and
UN
each x in \a b]
depends only on .
[a, b].
[CHAP.
INFINITE SERIES
228
a<x<b,
a^x^b,
11
such as
etc.
The domain
for which
SPECIAL TESTS
UNIFORM CONVERGENCE
for
SERIES
of
1.
then 1un{x)
is
Example:
\Un{x)\
(6)
2Af converges
can be
1,2,3, ...
5-^
n=
g Mn
(a)
[0,277-]
since
-j
-r and
converges.
This test supplies a sufficient but not a necessary condition for uniform
convergence, i.e. a series may be uniformly convergent even vi^hen the test cannot
be made to apply.
One may be led because of this test to believe that uniformly convergent
Hovi^ever, the two properseries must be absolutely convergent, and conversely.
without being
convergent
uniformly
be
can
series
i.e.
a
independent,
ties are
See Problems 30, 123.
(a)
is
(b)
Then the
Ui {x)
+ Ui{x) +
...+Un{x)\
< P
a^x^h
for
all
n>N.
series
CO
aiUi{x)
is
THEOREMS
If
an
If
to the
a2U2{x)
uniformly convergent in
on
]^a'u(x)
a^x^b.
ties possessed
Theorem
by sums
uniformly convergent,
it
has
many
of the proper-
6.
{Un{x)],
sum
TO
S{x) in
1,2,3,
[a, b],
if
CHAP.
INFINITE SERIES
11]
In particular
if Xo is in [a, b],
\imX'^n{x)
x-*x n = l
Theorem
If
to the
left
229
lim
M(a;)
y,Un{xo)
=i
n=lx-*XQ
hand
limits in case Xo is
..
an endpoint of
[a, b].
7.
n-
{Unix)},
sum
S{x) in
1,2,3,
[a, b],
.,
if
then
S{x)dx
21
<'^Unix)\-dx
Un(x)dx
(4)
"=l-'a
"
or
Unix)dx
(5)
Briefly,
term by term.
Theorem
If
and
8.
{Unix)},
if 2w(a;)
S'{x)
X<i^)
(6)
n=l
or
rx{tM^)} = tj-xMx)
(7)
This shows conditions under which a series can be differentiated term by term.
Theorems similar
= 1,2,3, ..
{Un{x)},
uniformly convergent in
Un{x)dx
which
is
POWER
SERIES.
series
[a, b],
For example,
if
then
Mm. Un{x)dx
(8)
7.
ao
aix
a^x^
...
^= <^n^"
(5)
power
series in x.
It is often
convenient
In general a power series converges for \x\ < R and diverges for |x| > R, where the
constant R is called the radius of convergence of the series. For |a;|=.B, the series may
or may not converge.
-R<x<R,
special cases
R=
and
R = ^
-><a:<>
(see
Problem
25).
can
arise.
In the first case the series conconverges for all x, sometimes written
speak of a convergent power series we shall
it
When we
R>0.
(9),
where x
is
replaced by (x
- a).
[CHAP.
INFINITE SERIES
230
POWER SERIES
THEOREMS
on
Theorem 9.
A power
within
11
and absolutely
in
lies entirely
of convergence.
its interval
Theorem 10.
A power
This follows at once from Theorem 9 and the theorems on uniformly convergent
The results can be extended to include endpoints of the
series on Pages 228 and 229.
interval of convergence by the following theorems.
Theorem
Abel's theorem.
11.
of
a power series converges up to and including an endpoint of its interval
this
include
to
far
as
so
extends
also
convergence
uniform
convergence, the interval of
See Problem 42.
endpoint.
When
Theorem
If
12.
converges at x
'^UnX"
= Xo which may
n=
lim\^anX-\ =
n=
x-*x
If xo is
left
|;llimaa;4
n=
2 aa;?
(10)
n=
\x-yx
J
or
in {10) according as Xo is a
x^xo-
of
This follows at once from Theorem 11 and Theorem 6 on the continuity of sums
uniformly convergent series.
OPERATIONS
with
POWER SERIES
we assume
that
all
power
some
interval.
Theorem
13.
Two power
mon
Theorem
"
series can be
14.
Two power
series, for
^= "^"
example
^^^
2j CnX"
^"'^"'
n=
11
where
Cn
aobn
CLlbn-l
Theorem
Cl2bn-2
common
dnbo
(H)
interval of convergence.
15.
If the
power
series
^anX"
is
of x.
quotient can be written as a power series which converges for sufficiently small values
CHAP.
INFINITE SERIES
11]
Theorem
231
16.
EXPANSION
of
FUNCTIONS
Suppose that
f{x)
in the closed interval
and
Then
in
as
f{x)
f{a)
i,
which
lies
f'(x),f"{x),
.,/<">(x) exist and are continuous
and that f" +(a;) exists in the open interval a<x<b.
previous chapters (see Pages 61 and 95)
+ na){x -a) +
where
derivatives
its
is
^"
^n
(^
(a^-)"^'
i)j
^^(a;-|)(a;-a)
between a and
/(a)
-K"
(^^)
/'(a)(a:
- a) +
(Lagrange's form)
(13)
(Cauchy's form)
(U)
x, is in
a^x^h
have seen
-v/e
POWER SERIES
in
If for all x
and
$ in [a, b]
we have
lim
i2
"-*"
/^
(a;
- a)^ +
^^(x -a)^+
...
0,
(15)
which is called the Taylor series or expansion of f{x). In case a = 0, it is often called
the Maclaurin series or expansion of f{x). For problems involving such
expansions see
also Chapter 6.
One might be tempted to believe that if all derivatives of f{x) exist at x = a, the
expansion (15) would be valid. This however is not necessarily the case, for although
one can then formally obtain the series on the right of (15), the resulting series may not
converge to f(x). For an example of this see Problem 104.
Precise conditions under which the series converges to f{x) are best obtained
by
of the theory of functions of a complex variable.
See Chapter 17.
means
x'^
smx
^-31+5!-
2.
cos^
3.
e^
4.
ln|l
5.
iln
1 + x
1-a;
-,
1.
a;|
-_+___+
-f + ^- J+
^+3-
cc-^+^-^+
X^
O'^
6.
tan-x
7.
(l
+ x)"
:.
X*
x^
X^
'7*5
X^"~^
...(_i)-x^^__
+
x^
2T+3T+
a;2-i
x''
71+ (-1)""^(2^+
_<^<,
...
+(^+
a;""'
-^<x<^
...(-l)n-.J+
...
-Kx^l
-l<x<l
T+--- +2^:ri
/v2n
/y7
1+PX+ ^^P^x^
^l
-~<^<~
-l^^^l
.(-1)-|^3T +
...
J>iv--L)---iP-n
ni
l)^
[CHAP.
INFINITE SERIES
232
This
is
(a)
If
(6)
If
25
(c)
If
l<p<0,
(d)
If
For
all
11
is
>
but
= 1,
is
^ 1.
a;
Ka;^!.
l<x<l.
-1 g
1< < 1.
if
a;
SPECIAL TOPICS
1.
Functions defined by series are often useful in applications and frequently arise
as solutions of differential equations. For example, the function defined by
x''
Jv{x)
x^
2^|-^ ~
2(2p
x'^
+ 2) ^2-4(2p + 2)(2p + 4) ~
^"'^^
^ (-l)"(a:/2)''+^"
n\{n + p)\
n=o
and
is
F{a,h;c;x)
is
a(a
a-
+ 17^ ^ +
aby
+ 1)6(& + 1)
,..,,
^^^^
+ l) ^ +
x{l - x)y" + (c - (a + 6 + \)x}y' ,
i.2.c(c
0.
many important
properties.
00
2.
Infinite series of
complex terms,
in particular
power
series of the
form
^=
""'^">
where
+ iy and
may
to real series.
Such power series converge for \z\ < R, i.e. interior to a circle of convergence
+ y^ = R2^ where R is the radius of convergence (if the series converges only
for 2 = 0, we say that the radius of convergence R is zero; if it converges for all z,
we say that the radius of convergence is infinite). On the boundary of this circle,
= R, the series may or may not converge, depending on the particular z.
i.e.
Note that for y = the circle of convergence reduces to the interval of conx^
I^l
vergence for real power series. Greater insight into the behavior of power series
is obtained by use of the theory of functions of a complex variable (see Chapter 17).
PC
3.
(aioa;
amy)
{a2oX^
+ anxy +
ao2y^)
^Unix,y) can
n=l
In particular,
we
(18)
using double subscripts for the constants. As for one variable, we can expand
suitable functions of x and y in such power series using results of Chapter 6,
Page 109 and showing that the remainder Rn-^0 as n^ =. In general, such power
series converge inside a rectangular region |x| < A, \y\ < B and possibly on the
boundary.
CHAP.
4.
INFINITE SERIES
11]
Double
Series.
Un
Let
Smn
Ui2 Uis
X 2 ^1
sum
be the
^ ^ Mp,
converges to the
limi
P,
rows and
lim Smn
S; otherwise
it
first
S,
we
diverges.
Let P
sum
already considered.
Infinite Products.
in the first
Definitions
numbers
of the
If there exists
5.
233
(1
we say
(1
00
(1
+ Uk),
or briefly n(l
+ Uk),
converges to P; otherwise
it
diverges.
to
Theorems about infinite products can (by taking logarithms), often be made
depend on theorems for infinite series. Thus, for example, we have the
Theorem.
is
6.
Summability.
If the
Let
sequence
^-g"^,
in the
converge absolutely
n terms of
of the first
summable
Si,
+ Uk)
-,
52t is
'^
Si,
Asymptotic
series.
1,
'S.Uk
....
we can apply
If the
converges to
The process
sw
0.
- +
...
+; +
...
,)
5+|| +
...
+^
^ = 0,1,2...
i20)
INFINITE SERIES
234
Rn (x) =
If
Sn (x), where
f{x)
then S{x)
is
an asymptotic expansion of
called
11
f{x) is given, is
lim x'^Rnix)
fix)
[CHAP.
f{x)
and we denote
this
by writing
S(x).
we may
begin to increase,
Solved Problems
CONVERGENCE
1.
111
(a)
its
DIVERGENCE
and
+ ^-^ +
of
Sn
- l)(2n
(2n-l)(2n
"
"^
+ l)
1)
(2to
~ I^T+Iy-
2 V2^r=^l
Since
lim S
series is
last, cancel
{1
lim
(&) find
Prove that
(l)^
Subtract:
iSn
lim S
Another method:
Let a
lim
(|)3
^2n-l
^\
+ iy
2M
17,
= f,
j^
+
1.
(f
=f
)"}
2,
converges and
(f)-
= I + (|)^ + (!)'+
=
(|)^ + (|) +
= | - (|)"+i
lim 2{1
sum
its
i_\
^ 1/
2
2n
\^
+ iy
is A.
sometimes called a telescoping series since the terms of Sn, other than the
Since
first
and
out in pairs.
|S
3.
and
'^**^"
iA_i+i_i+i_...+^^
2^1
(a)
^^o^^^rges
l)
Ml
The
^A2n-l){2n +
CONSTANTS
of
sum
"*"
2,
SERIES
+
+
(|)
)"+>
(f
or
2{1
+ #+ =
Hence by Problem
2^
26,
sum.
(!)
find its
(b)
:rT
Chapter
its
sum
(f)"}
sum
is
is 2.
a/{l
r) =
f/(l
diverges.
3,
the series
is
divergent.
f) =
2.
CHAP.
INFINITE SERIES
11]
lim m
follows
235
\/n
+l-
from Problem
14(c),
is
Now
COMPARISON TEST
5.
1, 2, 3,
Let
Sn
Ul
+ Ui+
Then
Sn
Thus Sn
is
2m converges.
We
Ui
+ 112 +
Thus
to
+ lu ^
lnn<n
iin
Also, since
+ Vn s T
t)
SO, r g
T.
OSSnST.
or
^^
prove that
test
(i
*)
(i
made
+i+i+
if
>
Chap.
(see
3)
diverges.
""'
= 1
_L(8terms)
and Vn be
s 1+
...
we can show
that
^+1 +
...
-^, where p
is
a constant, diverges
1.
^^ S^,
and
i)
Let
+V.
+ vi +
vi
and converges
Since
+ V2+
By methods
1
V1
ps
8.
if
1 + ^ + ^+... =
1
s
+
^
i
i
i + i = i
i
s i+i+ i+i = J
i + i + i +
+ l^+--- + T^ S Jg + J5 + j.+ ... +
have
7.
and
if
+Un, Tn
hm
Since 2v converges,
etc.
(i.e.
6.
6.
QUOTIENT TEST
and
O^Un^Vn, n =
If
0.
""*"
is
Chapter 3
'" ^
"^i2n3-r
we have
positive.
lim -^
If
^
=
^ = J,
converges.
constant
A # 0,
prove that
2tt
converges
By hypothesis, given
n = iV + 1, iV + 2,
>
choose an integer
such that
AT-I- 1 to
\~-a\<
I
-^ <
Summing from
we can
<=o
^- ^
<
'
or
(A~e)
%Vn ^
1 to
Un
^n
+ e)
for
all
(A
"2 Vn
n>N
Then
(1)
ilf -> =0),
(2)
[CHAP. 11
INFINITE SERIES
236
9.
(a)
For large
we have
n.
+ 2n
n"
^^^i 2^3-1'
'
^=
approximately
i.
'
*>
Taking
^-^in^
t^n
+ S'
^$=^
and
v.=^,
1.
= 42
2i;
2.^
is
^^J^|^
lim ^
Since
(a)
2m
diverges,
1/w
also diverges
by Problem
8.
obtain an appropriate
Note that the purpose of considering the behavior of Un for large n is^to
1/m.
v
taken
=
well
have
comparison series v. In the above we could just as
lim
Another method:
(6)
For large
Un
w,
+ yfn
2n^-i
lim
Another method:
lim
Since
nf ^'Z^^^ ^ =
.
^^
Page
1,
_w_l
- -^ - -^
approximately Vn
1v, ^ ^"2^
and
Then by Theorem
4.
converges.
(e)
lim yv^'f-^^^^^')
Then by Theorem
^(|^)
lim
with p
n"'(^^ =
= 3/2,
2 l/n
10.
Examine
(a)
Page
lim
^^ < ^ =
6(a) yields
i'
2 ^"
(o)
>
n=l
2 sin^(X"/
(^)
n=l
Then by Theorem
with p
converges.
(6)
For large
n, sin (l/n) is
approximately l/n.
/A
r
lim
hm n'smM -) =
,.
with p
= 3,
/sinOMV
<
f
rr-
1
1
INTEGRAL TEST
11.
We
From
^^ ^^^"^"^
diverges.
for convergence:
lim w^e-"'
1,
Then by Theorem
I"-
N = 1.
Integrating from
a;
= nto
= n+
l,
made
if iV
/(a;)
/(n)
using Property
7,
Page
1,2,3,...
81,
n+l
M+i
> 1.
we have
= /(n+1)
a5
f{x)dx
1,2,3,...
= 2,
the series
CHAP,
INFINITE SERIES
Summing from w =
M2
1,
Af
1 to
M3
Mj,
f{x)
Ma
+ M3 +
^^
Problem
is
'
M2
M_,
{1)
can be omitted.
we
equal to S,
is
we
unbounded,
see
from the
hand inequality
left
in {1) that
see
2m diverges.
in {1) that
complete.
geometrically
12. Illustrate
and
Ml
is
^^^^ '^^
ii^ J
exists
^^'^^ ^'^
J
+ M
dx
237
the
proof
in
11.
Geometrically, Mz
+ Ma +
+ m
the total
is
result
13.
(1)
of
Problem
11.
f^.p = constant;
(a)
~=
x~''dx
j
^i
P<^, j^
^^
P>'^>
j^
M'-'-l
p^ \
j_
^=
",
"_
Af'-"-!
i-p
'
-# = InM and
|;^;
(c)
{d)
| ne'
where p=l.
1-p
t~^;
(6)
lim InAf
series diverges.
",
series converges.
series diverges.
Thus the
^^^
^^
JirX
series converges if
" Jim i In
(05^
p>
and diverges
if
S 1.
=0
and the
series
diverges.
XM
.
^'^^
Jif. J,
'^'^
J!]|?L
-i^'^ir
^4""'
*""*">
= 4"
and the
series diverges.
Note that when the series converges, the value of the corresponding integral is not
(in
general) the same as the sum of the series. However, the
approximate sum of a series can often
be obtained quite accurately by using integrals. See Problem 70.
ce
14.
[CHAP.
INFINITE SERIES
238
Prove that
<
-r
2)
From Problem
11
it
.,
follows that
lim
M-foo = 2
00
i-e.
<
Since
+
I
<
lim
The required
result
is
-,
from which
lim
M-.00 n = l
1
2 ~TTT
<
-r
obtain, on adding
<
TT
< ^, we
:;
<
T^
J^ X'
M_>oo
00-1
-|
TT
< 77+
t2
4
^in^ + 1
.,
11
W2
-,
_i_
required.
to each side,
<
+T--
tt
therefore proved.
ALTERNATING SERIES
15.
any term
(a)
is
The sum
2M terms is
= (di - c(2) + {as-at) +
+
ai (a2 as) (a4 as)
of the series to
SjM
Therefore {S^m}
Also,
S2M + 1
it
lim a
0),
Thus the
(6)
is
('*2M-|-l
and
is
SiM
partial
+ a^M +
sums
2M + 2)
~ <*2M + +
2
16. (a)
Sam
lim Szm+i
of the series
2M
is
ttjM-i)
"-aM
at
lim Sum
and
made
{"'iM
lim
ct^M
+ 4)
O'lM
^ o
1"
2M +
~ ~{'''2M + 2~
O2M+2.
'"
by hypothesis,
(for,
S.
is
in stopping after
+ 3~"''^2M + 4)
lim ajM + i
terms
"" ''2M
(^BM + S
= S
which
lim Sim
Since
i-
in stopping after
we have
S2^S4^Se^S,^
0,
a^u)
(a2_2
follows that
S2M
(ffljM-i
converges,
'^aM
(b)
first
terms
+ s)
term which
is
('''2M
+2
'''2M
omitted.
is
+ 3) ""
('^BM
Find the
+4
"2M + 5)
maximum
"
'
'
error
made
in
(c) How
first 8 terms and the first 9 terms of the series,
of the series are needed in order to obtain an error which does not
many terms
The
series is
'"
'
'
'
2m
+ 1'
^+
|^
Since
-^
-prr
2n
+l
2n-l^
-^
If
and since
p,
^r
lim
then
n^2n-lr
0,
it
\u\
= - y,
a+i
CHAP.
INFINITE SERIES
11]
Use the
(6)
is
positive
Similarly, the first 9 terms are 1 negative and greater than or equal to -^,
i+ J - ^+ i-
fj-
"
tV
t^
+ iV
IS
i.e.
(c)
results of
239
M terms
after
than l/(2Af+l).
is less
M a 499.5.
from which
.001,
Thus at
To obtain
least 500
ABSOLUTE
17.
CONDITIONAL CONVERGENCE
and
+m +
Ml
we must show
converges,
\un\
+ Mj,
=
^
Sm + Tm
Since
|m|
and Tm
(Ml
m+
2|m2|
|m|
lim
Tu
lim
M^oo
must
18.
also exist
convergent.
that 5m converges.
iMil)
2|mi|
\ui\
is
\u2\
{U2
+
+
+
|M2|)
...
lim (Sm
Sm
+Tm Tm) =
lim (Sm
(Sm-
M-<
23
337r
is in
>
(Mj,
1,2,3,
.. .,
is
|m|)
follows that
it
Sm+Tm
is
Tu) exists.
lim (Sm
by hypothesis),
Tm)
Tm
lim
M-o<>
proved.
^^^^ - ^^^ +
^
jL
...
M-oa
]172^
Then
2|Mj|
for
0,
\uj.
^^^
O
....
absolute value less than or equal to the corresponding term of the series
which converges,
it
19.
^ {-'^y-'n
(a)
The
absolute convergence:
^ (-1)"-'
2
^(-l)"-i2
.^
,^. which
is
However,
all i
1,
if
and also
|m.|
lim a
^^
lim
-j^
(6)
The
is
and
n=2
is
cin+i
0.
|m+i|
= ^-^1-^,
it
is
series converges.
conditionally convergent.
n Wn
-
lim f _^^_
"- ./,
* In^ x
exists
[CHAP.
INFINITE SERIES
240
C
Hence
lim
M^ooJf
dx
a;
f^,
In^
du
lim
= t\
r^ - ^r^
In 2
In My
M_,\^ln2
ic
11
exists.
Thus the
p^^-
i^^'^^' **^"*
series converges.
Then
=2
In
Another method:
S'"-
(.+i)inMn+i) =
;nk
^^^
To examine
where
lim Un=
Since
(c)
^^
lim m ?^
that
""^
we must
absolute convergence
its
proceed
To show
On
show that
to
it suffices
0,
*"-'
^*
'
n-+oo
;nk
lim
|ttn|
lim
-j-
0.
RATIO TEST
20.
lim H-tL
n^ M Un
By
first
the series
= L<
hypothesis,
L<r<l.
U1
+ U2 + U3+
we can
< ru^
< r-M^ + <
< -M + 2 <
Mk + 2
<
u{r
we
\^i:-tl\
'^n
we can prove
where
z=
L =
that
if
lim
|mi|
L<
1,
1.
l^al
\us\
Then by the
Hiili
i,
>
the
series
2m diverges, while
<
<r<
consider
lim fetil
^
n-+oo
Similarly
+ r^ + r^+
test, since
(un+i/un)
r'u^
addition,
tt
lim
n^N,
all
r"M
Mn + 3
n* w
Then
M^ + s
By
We
non-negative.
1,
Mn + 1
etc.
is
Here M
convergence of
(a)
T n^e-"',
lim
{n
+ ly
n
n-+oo
1,
e"
lim ('^-lYe-^-'
<
'^-
11=1
'-^
(c)
Jl=l
^2.1
'( -r i
Then
n''e~"''.
lim
Since
{b)
n=l
_-(m2 + 2n + l)
liml2L+l)!^
n 6""
n-oij
lim (^^?^t^ ^)
n^x
lim e"^"-'
n-+oo
I'O
if
CHAP.
INFINITE SERIES
11]
(-l)"-'2"
Here u
(6)
Then
Mn+l
lim
(-1)2"+'
lim
Here
(c)
>
1,
(-l)"-'w
u.
(-l)"-'2"
Compare Prob.
(n
+ 1)^
19(c).
Then
.^' + l
n^| (-l)'(n + l)
lim
^|(n+ 1)^ + 1
lim|*Li|
2n'
+ lf
(n
Since 2
241
By using
test fails.
(-l)"-in
(^+l)(n-+l)
H^
is
seen to be convergent.
MISCELLANEOUS TESTS
22.
where
...
Here the
ratio test
Un+l
inapplicable, since
is
2|r|
or i|r|
test,
^^
2/3,
(6)
-2/3
V2\r\
if
if
Mis even
lim
Thus
y/lul]
if
\r\
Hence the
=
<
\r\
y/2\F\
^ /
=
li,
lim
]^)
n(l-
\r\
if
and
>
|>-|
is
odd
+ 3-6-9
ihn
li^
and diverges
(6),
1-4-7
=.
|^|
=
-
...
in case
(c).
+ l-4-7-..(3n-2)
3-6-9...(3n)
(||lj =
1.
However, by Raabe's
....
test,
=4
Ji _ fSniy\
24.
^-Y
2)
The
'
odd or
1).
lim 2"
(since
is
we have
I V\r'\
23.
depending on whether n
even.
Then
(a)
4/3.
n-.
Mn
+ f'^-^Y +
V2-4y
P^^\ =
lim
\^
Un
'
_\
('^'^'^
f l-S-5...{2n-l) \
V2-4-6; +
lim' ^^
n-V 2m
\J
-^\^
+^
+2
[
1.
2-4-6...(2n)
'
\2n + 2y
division,
/'2re+l
V2n + 2y
1
"
1
n
_^
'
5-4/w
4i='
+ 8i + 4
where
|c|
<P
INFINITE SERIES
242
SERIES
25.
of
(a)
/yin
11
FUNCTIONS
[CHAP,
00
;-^
\n 1 /y.2n
Assuming x^'Q
oo
(if
a;
(n
00
^<
\x\
series converges if
a;
If
a;
=3
= 3
n=i
\x\
(-1)"
on
\x\
If
1.
y^ =
1,
i.e.
3, the test
fails.
which diverges.
(-1)"-^
which converges.
3=i
3 S
of convergence is
+ l)'
"^
n=l3W
il!^ 3(w
y^ >
if
1
= ^2~
3n=lW
-I
2 tt"
a;"-'
and diverges
1,
CO
If
+ l)'3"+'
\n
we have
lim
n-*
m(^
00
a;"
lim
Then the
series converge?
< 3. The
a;
3 < < 3. At
a;
a;
= 3
conditionally.
(6)
Proceed as in part
lim
n-^
Then the
M < <
a;
(c)
Un
n\{x
'*"
w(a;
a)",
- 1)"
- 1)
2(3n
lim
(2n - 1)
n 1 ^2n
(-1)"-'*'
x^'
+ l)!
(2n
(2w
Then
1)
-,o(2n+l)(2w)(2TO-l)!
lim
'*""''"
(w+l)(a;-l)''+'
2+'(3n + 2)
(tt
=
Un
lim
n-+ 00
n^oo
The
For
a;
=3
Then the
x'a.
is infinite if
'
lim
"^
|a;
1|
[a;
i.e.
n=i
zero.
For X = 1 the
approach zero.
,
Then the
series
becomes
(2
l)!"'
x^
i.e.
lim
r;(2n+l)(2n)
'
lim (n
1)
|a;
is
a],
= a.
Then
x-1
l)(3w-l)(a:-l)
2n(3n + 2)
1| <
2,
x^
+ l)!(x-a)" +
n\(x a)"
(TO
lim
n-*oo
Thus the
_<.,
This limit
lim
.'
(-1)"
00
n-^oo
iA\
with m
(a)
a;-l
= 2
or
a!
=3
\x-\\
1| > 2.
and x = -1.
n=i
which
^3w -^r
1
la;
1|
<
2,
i.e.
2 <
a;
1 <
or
1 <
a;
<
3.
CHAP.
26.
INFINITE SERIES
11]
(,
243
2__i___
_|^(^|J,
^
ce
Then the
if
If
a;
If
or
If
a;
The
<
J-
diverges
1,
+ 21
= b-TT
la;
l^il >
1
if
^^1.-2.
^^
.^
If
and the
1,
test fails
a;
Thus the
(6)
r^^
series converges if
+2
]x
+2
^,,,^,
the series
is
|||| <
lim p^L\
M
-.
that
= -^
1,
where
^''^'^^
^'
and
a;
- -2,
i.e.
for
(a;
+ n)(x + n-ir
a;
g -i.
w
^^^^^e^'
"O^'^S^
1
(a;
we
see that if
Sn
a;
+ n){x + n-l)
- 1, -2,
?^ 0,
lim Sn
provided
1/a;,
n-*eo
Then the
+n
~n,
and
(l~-A\+(^
+ .+ ...+ =
MX
+ n-1
x
0,
L^+.
+n
-2 -3
-1,
'
ar
and
.,
its
sum
is
1/a;.
UNIFORM CONVERGENCE
27.
x{l
1:
Sum
of first
M terms
=
=
S,(x)
(l-x)
.T
-l-x'
=
- a;") =
If kl
<
1,
lim Sn{x)
If kl
>
1,
If
a;
1,
If
a;
-1, Sn{x)
S,(a;)
lim
and
(1
lim S(x)
(-1)" and
+
+
x(l
a;
1.
0.
Thus the
Method
2,
The
converges;
if
a!
= l.
series converges if
a;
= -1
a;
= 1,
i.e.
for
-1< S 1.
a;
test.
series converges if
Thus the
\x\<l and
\x\
If
<
a;^l and
1,
diverges
a;-^ (1
a;),
then
lim
|a;| > 1.
The test fails if \x\
series converges for -1 < i
if
Then the
f^l =
= 1.
g 1.
If
a;
lim
\x\.
= rthe
series
[CHAP.
INFINITE SERIES
244
28. Investigate
By Problem
^X-x",
SAx)
27,
We
in this interval.
S(x)
have
Rn(x)
--
e,
n>
by In
since division
But
if
\x\
of X, the series
In this case
<
(which
In
i,
\x\
(d)
<
la;|
<
negative since
is
|,
In
|a;|
and
In (i)
^ S S
a;
<
number by choosing
|a;|
find
N dependent
>
or
j^
inequality.
^) reverses the sense of the
Thus
since
is
independent
j^ = N,
> j^?^ S
is
also
sufficiently close to 1.
1 < <
a;
sum
all
^^
Thus no
and
exists
it
1.
function
.99,
S(x)
it
limSn(x)
of
a;
Oga;<l,
Thus
points in
x'
^ ^ 1.
interval
a;
<\ne
If
we can
>
-J.
If
- a;")
(1
'^
29.
any
1^ > 1^ = ^-
and >
In (^)
n\T:i\x\
\x\
Now
positive
(e)
when
is
\x\
|a;|
la;"l
uniformly convergent in
(c)
Sn(x)
the interval
in
0^x<2.
(e)
-^<'^<h
i*
(6)
S{x)
The
on
{d)
ln ()
"""
-l<x<l,
11
Sn(a;)
l,
S(x)
a;
<
S{x)
P
I
Urn Sn(x)
and S(x)
^*
if a;
= ^1
"^
=
"'
lim
(1
- a;") ^
1.
0.
and S(x)
is
discontinuous at
a;
but continuous at
all
other
1.
function
In Problem 34 it is shown that if a series is uniformly convergent in an interval, the sum
in an
continuous
not
function
is
sum
if
the
that
follows
It
interval.
in
the
continuous
S(x) must be
This fact is often used to demonstrate the
interval, the series cannot be uniformly convergent.
non-uniform convergence of a series (or sequence).
30. Investigate
Suppose xv^Q.
Prob. 25, Chap. 3).
Then the
x^
^(-^
l-l/(l +
x-)
^^
"^
"
"^
"
'
1/(1
+ x^)
(1
+ x^Y
'^
whose sum
'"
is
'
(see
CHAP.
11]
INFINITE SERIES
U x~0
sum
Since
the
lim S{x)
n terms
of the first
S(x)
S(0),
S(0)
is
hence
0;
discontinuous at
is
245
;.
S(0)
= 0.
lim S(0)
0.
The^by Problem
is
(absolutely)
Lludes
"
= o!
WEIERSTRASS
TEST
31.
'''''^
test,
^^^^
i.e.
^^^
if
'=""''^''^^''
?onfergent"'^^"*''
l(^)i
\u,,{x)
u,,{x)+
n terms
R^x) = u.,4x)+uMx) +
is
Now
+ 1-^,,(^)| +
g M + + M+. + ...
But M+, + M,,+ ... can be made less than e
by choosing n>N, since 2M, converges. Since AT is
clearly mdependent of. we have
\RAx)\ < e for n > N, and the series is uniformly
convergent The
absolute convergence follows at once from
the comparison test.
...\
==
|Mn+i(:r)|
...
32.
cosnx
y2?i^
(a)
|5J**
(a)
w*
^-
Then
By
all
converges.
sin
nx
\n^
for
n-
2M
^sinnx
v^sinnx
n
-^i
1
^
+ x-\=
I
J^
n-'
= Mt g
-1 g
_1_.
+ x^
a;
i,
1),
the series
j.g.
|a;|
is
uniformly
1.
it/=^, we
-1 g s i by the M
a;
The
see that
2M,
test.
^ converges.
Then by the
a;.
enough,
=4>
choosing
33. If
(a)
^iri'
M=^,
^n'*
test.
we cannot
all
||1|
.^
,
^""f
'
x by the
case and
this
(see,
all
in this interval,
(d)
'
\~^^\ =
in
,
^
'
For
^"^
since
(6)
X"
.^ ^"
n^iM-"^
(h\
i.e.
)a|
<
Jim ax^
for
n>N.
\a.x'\
\x\
it
converges
(a)
where
|x.|
|xi|,
\a.x2\
<
absolutely
< |xo|.
by choosing n large
Then
|alki"
<
5t^
[CHAP.
INFINITE SERIES
246
11
Since the last series in {1) converges for |a;| < |xo], it follows by the comparison test that
the first series converges, i.e. the given series is absolutely convergent.
(b)
M,=r^-
Let
Then 2M converges
since
<
la;i|
As
|a;o|.
part
in
(a),
|ana;"|
< M
for
Ixol
jxj
jXiL
THEOREMS
34.
is
S{x)
where we choose h
continuous in
is
+ Rr.{x),
S{x)
S{x
S{x
so that
+ h)-S{x)
so that both x
h)
Sn(x
+ h) -
S^(x
and x
h)
is
<
\Rn(x)\
Then from
{1), (2)
\S(x
35.
\Sn{x
\h\
<
S,
interval
[a, b].
h)
Sn(x)
lie in [a, b]
(if
+ Rn{x +
+
Rn(x
= b,
and thus
h)
h)
and
+ h)-
e/3
S{x)
<
whenever
c/3
and
\Rn{x
\h\
we can
uniformly convergent,
+ h)\ <
R,(x)
(1)
for
its
Page 228.
6,
Now
uniformly convergent.
is
UNIFORM CONVERGENCE
on
Prove Theorem
We
2 a,. a;"
test,
<
also be continuous.
(2)
choose
for
e/3
must
it
< 0).
so that
n> N
(3)
(3),
\Sn(X+h)-Sn{.x)\
=S
S{X)\
is
\Rn{x
h)\
<
\Rn{x)\
established.
Prove Theorem
7,
Page 229.
If a function
continuous,
is
continuous in [a,b],
its integral
Then
exists.
S{x)
S{x)dx
Rn{x)dx
r"
I
S{x)dx
r'
Sn(x)dx
r'
I
Rn{x)dx
can be made arbitrarily small by choosing n large enough. This, however, follows at once, since by
independent of x
the uniform convergence of the series we can make |i?(x)| < e/(5 ct) for n>
in [a, b], and so
This
is
r"
Jj a
Rn{x)dx\
I
r"
^I a
\R(x)\dx
r"
<
.>'a
r^^^
^
^
"
S(x)dx
lim
S{x)dx
or
lim
Sn{x) dx
lim Sn{x)} dx
CHAP. H]
36.
INFINITE SERIES
Prove Theorem
Let
g(x)
8,
247
Page 229.
2^m:(x).
Since,
by hypothesis, this
series converges
uniformly in
[a,b],
we can
g{x)dx
u:{x)dx
u{x)
{u.(x)~uJa)}
Un{a)
S{a)
*^- ^'^-
^^^^
S{x)
00
because, by hypothesis,
prove?frt"hiri
37.
Let Sn{x)
2^ m^x)
converges to S{x) in
nxe-^^, n =
(a)
Determine whether
[b)
(a)
j^
SWdx =
1, 2, 3,
a;
'
''^'
-(^)
'(^).
which
1.
f &(x)daj =
lim
(a).
_ a;e--^dx
lim
S(aj)
''^'
X^^^^^'^ =
[a, 6].
lim S(a;)
lim
= 1(1-6-).
-ie--^|;
S(a;)da;
nxe"^ =
lim
0,
s(ic
(cc)
1.(1
- g-")
whether x
dx
Then
= Oor 0<a;Si.
Then
^0
It follows that
rs.(x)da;^
lim
"
(6)
f limS(*)da;,
^o ""*"
i.e.
convevgeumforr>ily to
j^
38.
Let
f{x)
We
all
a;,
= J?H.!^.
have
^Hl^^
in particular
^.
f{x)dx
Prove that
C0SW;7
1_
2^
-fi(27Z-l)
Thus
<F+^+F+-)
^|(2^
1)^
[CHAP.
INFINITE SERIES
248
11
POWER SERIES
00
39.
2= "^"
series of derivatives
Tl
OS
Let
we can
i?
>
choose
so that
|a|
<
< R.
\xor
series
InunX"'^
2w
however,
|a;|>i2,
lim "
= 2
can for
re |al |a;|"-'
n> N
'
,
,
\xo\
n> N.
for
rr
<
Let
all
and thus
lim oa;""'
how
close
|a;o|
so that
0,
for
test,
|a;|
<
|a!o|
< B.
R).
is to
converge.
Thus
is
40. Illustrate
Un + 1
lim
-+00
The
may
series of derivatives
\x\
may
or
3.
o
^2 'o
=in
At = 3
m^
,.
lim
n-*
00
3(w
By Problem
n^
3"
\x\
% 3"
3 ^
a;
i.e.
<
3.
R = S,
Note that the result of Problem 39 can also be proved by the ratio
when the
22.
Prove that
in
its
interval of convergence a
(a)
(b)
(c)
We
(c)
'
^
.=1
The two
same interval of convergence.
(b)
(a)
+ lf'
series of derivatives is
fi
41.
= i2.
a;"
2nx
Problem
|a;|
X"
series
(n+l)''3""
<
power
series
f{x),
consider the power series 2aa;", although analogous results hold for 2a(a;
This follows from Problems 33, 34 and the fact that each term " of the series
This follows from Problems 33, 35 and the fact that each term anX" of the series
and thus integrable.
is
is
a)".
continuous.
continuous
39, the series of derivatives of a power series always converges within the interval
of convergence of the original power series and therefore is uniformly convergent within this
interval. Thus the required result follows from Problems 33 and 36.
From Problem
If a power series converges at one (or both) end points of the interval of convergence, it is
possible to establish (a) and (6) to include the end point (or end points). See Problem 42.
^^^^- "]
INFINITE SERIES
^^'
if
249
rrZ,?'^''
^^^.r'",?^*
of convergence,
then the interval
For
we assume
power
the
series to be
.'
interval of convergence at a; =
1, so that the series surely converges
^"'"^^''^^^ ^'^'^
for
show that the series converges uniformly in
this interval
its
<
<
=r=
a;
TV,o we
..
Then
must
1
1.
Let
Now
Hence for
a;
<
{i2
R^,)xn
fin*"
'
(;''
...
i)
i,
lR,{x)l
smce (l-a;)a
=
=
=
Rn(x)
|fil
+ + a;2 + a;'+...) = 1
= l, |fi (a;)| = |5.j < ,
a;
(if
Also, for x
re.uire'd^^elKlWs/"'^
^"
'^
(1
-^)(|fi,,|
for
^' "'^'^
''
<
a;
>
1^,,,,^
,^^,,^.
...J
^^^
1)
^^ ^"'^^^^"^^'^^ >*
*^
-'-
of . in
1.
and the
to be
lim
a^x"
Sx g 1.
*.
Prove that
tan-i x
vergent in -1
(6)
Provethat
(a)
By Problem
;r
x-~^ + ^-^-u
^
7 +
1.
r,
where
the
~= i-i + i-.! +
25 of Chapter
3,
Y+^
Integrating from
to
with r
= 1~
where
a;,
= -x'
x'
-K x <
and a
= 1, we
+ x'-x'+
1,
have
...
-i<a;<i
^j^
yields
.-^ ^-^+...
7^
= ^--'- =
+ 5
./o +
(:2)
3
using Problems 33 and 35.
Since the series on the right of te)
- +1 it 4-^11
converses ^''
for r '
.,
t,
..,
"^
series is uniformly convergent
-}' ^* '"""^^ by Problem 42 that the
in
y
1 = x s 1 and represents tan-'a; In this
interval.
flfx^
m -l-^x^TlZ
(6)
By Problem
^lim
43 and part
tan-x
(a),
^lim
we have
(x
-^+ f - ^+
..
.^
:^
357^
,
45.
[CHAP.
INFINITE SERIES
250
We
have
1-^ dx
Evaluate
e"
+ 7T+"^+'
+ -5T
4!
5!
3!
-57
2!
ThenifM = -x^
Thus
we
zi
X'
= l-=^+f!--f!+ir-i!-+---'
e-^
=
-
X^
x"
X*
^~2T3!
1
^^
_
-
_
~
=
Note that the error made
fifth term,
i.e.
less
than 0.001
1- +
adding the
Problem
(see
_^__ _
I'
'"L
0.03333
9'5!
first
0.00595
0.00092
0.862
is less
than the
15).
(i 6),
x^y"
S a; S 1,
1_ + ^^
7'4!
5'3!
0.16666
MISCELLANEOUS PROBLEMS
46. Prove that y ^ Mx) defined by
The
<
*=
9-5!
7-4!
5-3!
3-2!
in
in particular,
so,
_x^ _ _x^
3-2!
<
and
_^^
-=
4!"^ 5!
X'X-e-2
-2
11
Page
xy'
(x^
- p^)y =
be
x [see Problem 106(a)]. Since a power series can
x,
for
all
interval of convergence, we have
differentiated
its
.ro 2''+''n!(M
^'
,?o
+ p)!
+ 2n)(p + 2w--l)x^+^
2''+^"m!(TO
" (-1)- (p
y"
+ p)!
n=
Then,
{x^-p^)y
_
*^'
a;Y'
i^gp+^'MKn + p)!
+ 2w) x'^'"
+ p)!
" (-1)" (p + 2w)(p + 2w - 1) x''+"'
2^
2''+'"i!(n + p)!
2''+'"'w!(M
00
X2/'
(x'-p')y
J\njj.p + 2n + 2
S^a'+^-wKw +
"*"
P)!
(-1)" [-p'
"^
(p
+ 2w) +
(-i)"a;''+''+^
?o 2-+^n!(M
.?! 2"+'"-" (n
"
2"+='"
+ p)!
(_l)-i
+ p)!
" (-1)" (p
.=0
Adding,
xj/"
.r2''"M!(n
- 1)
"*"
(-1)" [4ji(m
(n
+ 2ffl)(p + 2n - D]
x'-^"
?)!
+ p)] x"'^*'
+ p)l
2^*'"-n]{n
a;p+2''
!
(p
n !(. +
- 1 + p)
(-l)Mx''+^'
"
-r,
.|
( 1)" 4x''+'"
2"+=''
(n-
1)! (m
+ p- 1)!
(-l)Mx'-^"
.=1 2''+"(n-l)!(w
+ P-l)!
CHAP.
INFINITE SERIES
11]
251
-n
so
47.
Since
lim
lim
(n
w
-^<'^,
i.e.
<
\z\
For
- 3,
|z|
1)'
Thus the
z^ix
e''
cos X
Similarly,
49.
Prove that
lim(
n--oo \
Letting f{x)
==
2,
i sin
e"'^
cos
Consider
S..,
a;
sin
respect to x from w to m
'
M
I.e.
Sn+i
Sn
/"
*
n fi^of
to
0.577215.
1^1
Z
-g-,
2A
|z[
absolutely
sin x
we have
,
..
+1 -
inw
ft
we
11,
by
exists.
find
n,
I = l+i + | + i+-+^-ln ^ 1
+1 + 1 + 1+ ... +l_in jg bounded by
1,
in
(H1\
By
and
1.
-^ g 1 g 1
with
we have
^in
so that S.
is
monotonic decreasing.
infinite
product Yl{l
is
converges
''"'
equation
is
= 3.
'^
.
.
50.
+1
S 0,
^%3
' + ^'
/'i
\ ^
Y x' x' \
(^l---...j
+ ,^^__+^_...-)
The
a;.
Problem
= -1^ -
_)_
(i),
'
3,
circle
cos X
^ + ~+...,
+|^ + J-O + i+
^
1/x in
2 jSJ^
is
and on the
+ ^++ -
ix+
+ tx+
+
g/
3.
e'^l +
in
^im
= 3.
\z\
e"
Letting
>
\z\
3"
3,
48.
z
5 -^r-^
for
(J)
if
+x S
e'
tor
x>0,
equal
'^u^ converges.
k=
4,
y, is
rational or not.
so that
Pn
- n(l + Wfc)
(1
+ Wi)(l + M2)
(1+ M)
e"i
e"i+"z +
"-+%
fc=i
is
51.
[CHAP.
INFINITE SERIES
252
Then
Si
sums
of partial
Si + S,
2
+ _
- ^
1,
we
52. (a)
1 1/2
^ %/(2i
- 1)
Prove that
if
1, 0, 1, 0, 1, 0, ...
is
if
is
if
n.
is
even
odd
>
'
^^^^
and p >
= 1^
jj^^
p{p
+ l)
to 1/2.
+ S2 + S, _
~
3
S1
+ +
_
~
3
1, |,
2
'
'
'
|, f. |, i-
'
'
follows.
0,
g
J-^t"
(b)
Use
i.e.
(a)
(a) to
Jl
f?
.. p(f>
+ l)---(p+w-l) l
prove that
the, left.
we have
Integrating by parts,
/.
is
^du
lim
Hm
e-'u-'-du
i(-e-")(Jt-'')''
(-6-")(-pM-''-')dMl
J"
dw
^
p/p+i
a:"
a;
Similarly,
e
~
- ^ri
/p+i
(P
+ 1) ^"^^
so that
a;
By continuing
(6)
Let
S(a;)
Then
in this
e-|- "
Rn(x)
\Bn{x)\
p(p
f(x)
^p+.
Sn(x)
"
result follows.
("l)
^i^J"
(-l)+p(p
+ l)...{p + n) r -If^dw
+ 1) (? + n)
e-'dw
J
e-'dzt
1.
Thus
z^r^du.
p(p
+ l)...(p + w) f
p(p
+ l)---(p + w)
_p+ll+l
since
Now
^^^dM
CHAP.
INFINITE SERIES
11]
lim \x^BAx)\
and
follows that
it
B. (a;)
lirn^
Mn
|Mn+l|
,.
lltTl
_
~
^^^ *^
+ ^]---iP + ^)
0.
diverges for
'"*^
result is proved.
^^^
PiP
lim
253
we have
for
all fixed x,
all
Supplementary Problems
CONVERGENCE
53.
and
DIVERGENCE
(a)
sum.
of
i:
SERIES
3-7 ^ 7-11
Ans.
of
CONSTANTS
11-15
...
i-
{4n-l){4n
.-f.
+ 3)
<=<'^''^^Ses
and
1/12
(6)
54.
55.
If
56.
57.
2m and
fallacy:
1)
(1
^^et
and
(fP
- 1 + 1 - 1 + 1 - 1 + ...
=0. Hence 1 = 0.
f- 1) +
(1
(f)'
5(|)"
5(w
+ i;)
converges to
A +B.
diverges.
Then
5 =
(1
- 1) - (1 - i) -
...
and
QUOTIENT TEST
and
Ans.
59.
- 1) +
COMPARISON TEST
58.
converge to
2t;
(a) conv.,
(6) div.,
(c) div.,
(d) conv.,
(a)
^
^w^
(e) div.,
5n
= I w(m2
(/)
-2
1)^'2
lo;
'
conv.
60.
61.
.?.
converges
test to
if
62.
63.
p>
and diverges
(a) conv.,
(6) div.,
1,
(6)
Page
^^IlpL
diverges,
225.
'
Ans.
if
(c)
+QT1-1
('^^It' wiV.tan-MW),
l/-
In~n"
lon'
'^^-
conv.,
converges.
(6)
prove that
and
(6)
\ 2 +
..
,-fJ
/I
(c) div.,
/S\
/\
(0)
(d) div.
11^;^,
n(l +
fi
(d)
l.sinMlM).
(c)
div.
64.
If
65.
(a)
0.
2 -^-
INTEGRAL TEST
>
^"'""'.
67
Prove that
68.
Prove that
(a)
Ans.
1?
(a)
Ams.
(a) div.,
where p
./
=2 %(ln n)"
(6) conv.,
is
J:^ii(hiW
(c)
a constant,
(a)
converges
() div.,
if
_
'
-iS^'
^^'
='
(/)
> 1 and
'=^
V^
'
'^
div.
(ft)
diverges
if
i.
2 :;^ < J-
<
^^a^^TTTi'
(&)
(a) div.
^^^
69.
11
n-*co
66.
[CHAP.
INFINITE SERIES
254
Arts.
70.
(a)
(b)
Provethat
Use
(a) to
+ V^ =
+ i ^ VT + V2 + Vs +
estimate the value ofVl + V2 + V3+--+
Show how
(c)
(b)
671.5
2
=2
'"
I-
maximum
error.
+ a/IOO
can be improved by estimating, for example, V^O + -/ll +
accuracy.
of
degree
desired
some
to
computed
/2
+
+
+ \/9
(~1)"V^.
ln%
(a)
Aws.
%^^
(a) conv.,
,,-,,
(6) conv.,
(c) div.,
1 ^3^^,
(^)
(d) conv.,
(e)
(d)
What is the largest absolute error made in approximating the sum of the
Atis. 1/192
the sum of the first 5 terms?
which must be taken in order that 3
(6) What is the least number of terms
Prove that
(b)
How many
accuracy?
74.
and
S -
^+
23
+ 33+""
""
i^
2'
3'
1)
^^
/
in order to calculate
CONDITIONAL CONVERGENCE
(0
Prove that
Ams"
(1)"
2^ 2"(n
'
<a)^^^
75
series
Ans. 8 terms
(a)
ABSOLUTE
("D" sin--,
div.
(a,)
will result?
73.
"
r_1^+
^2;^,^^^.
72.
(6)
4.5
ALTERNATING SERIES
71.
the accuracy in
"'
|n='^
=i
(6)
TT'"?
a;^ + n^
2^
cond. conv.,
(c)
(e)2^sinA,
all
real
a;
(e)
and
abs. conv.,
a.
(/)
abs. conv.
CHAP.
+ 1^ 1 +
If 1
76.
INFINITE SERIES
11]
!+ ^=|S.
Take 1/2
[Hint:
term
255
and write it as
Note that S = In 2, as shown
A +
in
0-J +
Problem
+ +a 4 + i + l-J- + la + JU
lY
5
+ J+.--;
96.]
Prove that the terms of an absolutely convergent series can always be rearranged
without altering
the sum.
77.
RATIO TEST
Test for convergence:
78.
Ans.
(a)
79.
Show that
80.
Prove that
(c) div.,
conv. (abs.),
(d)
(e) div.
2 ~ converges
and
lim
(6)
n-+oo
0.
'W.
MISCELLANEOUS TESTS
81.
82.
83.
|+
test to
(|)=
84.
If a, b
85.
work Problems
(i )^
(a)
(|)4
86.
of
if
-a >
d,
and diverges
(e).
converges.
2-5
,,,2^2-5
2
r^^
2-5-8
rf)
""
FUNCTIONS
87.
'^
and
a{a
a(g
Aws.
(|)
(d)
(c),
-^'^'^
'^''^
"^
converges
1-4
1-4 -J?
1
^
+ ^~^
+ illll
+
a^
a
SERIES
(ly
78(a),
()
-l^si,
Prove that
(6)
series:
i^^i,^..
~l<xS3,
^'^if-^^Ql^ ^^
(e)
all
x^O,
converges for
.|,.-(fS)". <.)i;.^=n
(d)
x>0,
-1 g
<
(e)
a;
S 0.
1.
UNIFORM CONVERGENCE
88.
By
[Hint:
1
(n-l)x]{l
+ nx\
sum
is
S.{,x)
+ nx''
^'^'
''"''^'"^^"*
^thJ^nU^ll
"
^"y
^"^^"^^^
^^^"^ mcWAes
a;
= 0;
[CHAP.
INFINITE SERIES
256
30 directly by
obtaining Sn(x).
89.
Work Problem
90.
Investigate by any method the convergence and uniform convergence of the series:
Ans.
91.
If
F(x)
(a)
Fix)
is
^^,
first
3;
(a.)
>
all x.
(c)
conv. for
x^O;
0.
prove that:
all x,
(6)
lim F(x)
92.
Provethat
Q^ ^ S^ + ^-f^
93.
Prove that
Fix)
94.
95.
Provethat
has derivatives of
all
-,
= j^f^. =
u^x)
1.2,3,
F'(x)
(c)
0,
<^x
^"!^"'
n=i Sinn fiTT
dx
'^^
r'
J^ j^^^j^.
lim
(6)
x^O,
continuous for
11
^^
continuous everywhere.
is
0.
.. .,
6-'
POWER SERIES
(a)
Prove that
ln(l
(6)
Prove that
In 2
[Hint:
Prove that
98.
Evaluate
Ans.
+ a5) = ^~Y'^Y~'4"^
= 1- ^ + i-i+
^-^ = 1 - x + x^ - +
97.
sm"' x
(a) 0.461,
(6)
a.
f'^e-^dx,
(a)
X^
X^
96.
a''
a.
s"
j^'
(6)
3 ^'
2^
4.
~5
xi
7
^
,
100.
101.
By
102.
Showthat
103.
Let
exists
/(x)
4,
[Hint: See
x'
tanh-;
(6)
ln(x
but that
j^"'""
it
_i
'
''
-^
<
-
J.
<
-
31x
4x*
r^-r^v
V^ + l)
a;#0
eT"^"/'
2 sin
a;
cos x
38,
Chapter
sin 2x.
^ ^^<r ,
-<*'<-
-1< x < 1
^+^ + f+--
-2
1 x
3"+
x''
1-3-5
x'
4.]
-1
2^ J" 2^T^T
xt^O.
does not converge to the given function for any
Problem
1
^
= (^1-2T + T!
(a)
'
e-"
integrate.]
0.486
99.
104.
1-3-5
2-4'6
Lll|^^dx
and
<
=
a,
a;
1
i
corresponding to /(x)
CHAP.
105.
INFINITE SERIES
11]
257
Prove that
^^
()
+ .)r
{ln(l
(6)
-(l
+ (l+| + |)=c' -
+!)''
for-l<a.<l
MISCELLANEOUS PROBLEMS
Prove that the series for
106.
Mx)
converges
for all x,
(a)
(h)
interval.
107.
Prove that
{a)-^{Mx)} = -J,(x),
= ~Jp(x) -
(c)Jp+i(x)
108.
109.
Prove that
eV4'-'/"
Write the
[Hint:
\z\
111. (a) If
2^
Suppose that
If I,
113.
114.
r^r^n;-?'-is
expansion
=1
lEi
J-^(x)
0,-1,-2,
.,
w=
(-D'Jnix),
prove that
1,2,3, ...
108.]
for
all
v'R =
(c)
common
in the
Use
to
(a)
interval of convergence
show that
if
\x\
<R
L.
L<1
or
Use
(c)
How
[Hint:
<?iir>i
thot
such
that
2 TTT^
criterion,
->'"
p/^'^i'i"'^
F(-p,l;l.,-x)
(1
l<j
|&p
Show
that S'ix)
<:
limits,
converge
i,
whenever
is that,
"lai,,
p>N
-.
e ^
> u,
> iV, where
given any
and g
converges.
2-
Page
(fw 282)
_ S,| ^
<
>1
22.
diverges'
43.]
() is
!.|
<
1,
(6) i,
dlvg,l
+ .)..
f
^
the
"^Jf^
(a)
on
"
>'l'
Problem
deDendinff
"''^"'""'"^
j_
a.
(a) to
""'
f^lrf
N>Q
find
<?-+
(h)
i^
we can
B>0,
where
115. (a)
118.
x^ J,.,{x),
t^
e"'' e-'',
^^
6a;"
**
exists, the
the
112.
J^ix),
^p(a;)
left side as
aa;
2^^:p2J^
1.
JMx) =
(6)
J^-i{x).
result of
= -Mx),
J-,(x)
(a)
j-{x^ J,(x)}
{h)
S(x)
l
xS{x)
-^ +
1*3
and
'^^
j-
l'3-5 ^
solve.]
'
Ans.
..
e'^'^
Jo
e-'^'^dx
[CHAP.
INFINITE SERIES
258
11
Prove that
at
119.
^'"^'^
Prove that
Use the
[Hint:
-4!"^
is
n=i
2*
on Page 228.
121.
~ ^V^~^'^2^-2!'5"2-3!-7"^
lT3""'"l'35"^ l-3-5-7'^
Page
Dirichlet test,
and Problem
228,
Chapter
94,
0, ff, 2?r,
1.]
[Hint:
123.
Prove that
124.
Prove that
125. If
a;
converges
2
x^
n+r-r1 -^ + ^- ^+
4
7
10
^^
= ye",
^1
prove that
1/
11=1
126.
uniformly for
3y^
("1)" '^"
ni
Let
^
e''
Show that
(a)
(6)
Bi
= -^,
- ^
129.
ft
= 1,
^2
= 0,
Bz
Bi
= -^,
= -(coth|-2^
^
^:"i
1,2,3,...
it is
given by
" (-l)'-'n"-^e-
+^^
+ ^^+ .
3
2
= 1+ Bix
1
<x^ lie.
-1/e
for
x"
_
127.
all x,
^ + |-ln2
3
= X-1
e"'^
in Taylor's theorem.]
1).
= 0,
Bs
Be
=
.
where B"
.,
is
Be.
=h
(6)
(a)
to
,x
x'
B2n
(a)
cothos
^+ 3-^5+
(&)
coto;
^-3-45
(-1)
[Hint:
For
for
use esc
(d)
130.
Prove that
131.
Use the
132.
Prove that
(a)
fl
a;
f1
cota;
+ -j
(2a;)''"
+-(2^I)Tir+
-(2;oiir+
(6)
replace x by ix in
for
(a);
(c)
use
tana;
cot a;
2 cot 2a;;
tana;/2.]
converges.
(l
+-)
diverges.
00
n=
133. (a)
Prove that
(1
- *").
where
< m <
1,
converges
if
and only
if
2m converges.
11
"^
converges to i.
(6)
Evaluate the
infinite
product in
(a)
134.
135.
is
regular.
is
C-1
[Hint:
summable
to zero.
3.]
to 2 decimal
CHAP.
136.
137.
INFINITE SERIES
11]
series
is
+ 2x + 3x' + 'ix^+
...
Abel summable to S
called
+ nx"-' +
S =
if
(-l)"(n
^
2,
(-!)"(
n=
th\
(6)
1)
~+^-^
l){n
is
+ 2)
'-
IS
>
Abel, summable to
1 or
140.
Use
(a) to
On*"
for
|a;|
< 1.
Prove that
exists.
J^
CO
2 (^.
1/8.
.|
+ ^y
where p
is
1 respectively.
= l-J- + 2!_3!
f!Zd^
(6)
- x^
00
138.
lim
converges to 1/(1
X-+1- n =
00
(a.)
259
prove that
C ?^du
M
J.
(-ir-'(n-l)!
r^-".,
i--l+21_31 ^
x
a;=
a;
a;*
Prove that
f'^iHiidw
M
^SlULdu
X
141. If /(a;)
9l^{l_2}_,4l^_
a;'
x*
^
\
2= ^^^
% 1
expansion
?\
sinx/l_3! 5!
X \x
x^'^x'
'")
sing /
X V;
4!
x^"*"^
'")
-, prove that
2!
f{x)dx
chapter 12
Improper Integrals
DEFINITION
of
IMPROPER INTEGRAL
an
dx
f{x)
is called
an improper integral
if
(1)
(2)
a=-a>
or &
or both,
<
i.e.
fix) is unbounded at
larities of f{x).
a^x^b.
...
is
infinite,
called singu-
Example
2-
-^
Vx
J,
4:
is
is
first kind.
00
Example
dx
dx
is
^^^dx
is
sm
we formulate
IMPROPER INTEGRALS
Let
f{x)
of the
tests for
FIRST KIND
f f{x)dx
The
series
ztn
where Un
infinite series.
Similarly,
We
we
Note that
while
f{n),
often write
lim
f{x)dx
a^x^h. Then we
define
is called
finite interval
f{x)dx
in place of b in {1).
define
f{x)dx
lim
260
f{x)dx
(2)
CHAP.
IMPROPER INTEGRALS
12]
261
and call the integral on the left convergent or divergent according as the
limit on the
right does or does not exist.
Examplel:
Example
2:
cos
exist,
= J
lim
cos X dx
we
In like manner,
da;
a;
Hm
f"^
cos
"
(^
da;
a;
l)
=
lim
so that
(sin
?^
converges to
1.
sin a).
divergent.
is
define
/^^
y^^O
fix)dx
-^CO
j_^f(x)dx +
f{x)dx
(3)
where
t>0 and
diverges
if i
S 0.
FIRST KIND
of the
e'^da;,
where
a constant, converges
is
series if r = e"'
so that e~'^==r^.
~,
verges
> 1 and
if
diverges
CONVERGENCE TESTS
for
if
g 1.
where p
is
IMPROPER INTEGRALS
a>0,
a constant and
of the
con-
series.
FIRST KIND
The following tests are given for cases where an integration limit
is oo.
Similar
tests exist where an integration limit is -=o (a change
of variable x = -i/ then makes
integration limit
Comparison
(a)
verges.
(6)
Then
if
Since
Divergence.
f{x) is
test
Convergence.
Example:
the
continuous
=o).
all
0^f{x)^g{x) for
^ ^ = ^^
and
all
f e-dx
a;
S a,
C f{x)dx
converges,
that
con-
also converges.
f^
that
j'" g{x)dx
also converges.
C g{x)dx
diverges.
also diverges.
Example:
Since
3^
diverges.
>i
for
a;
2 and
f^
"^2
f^
da;
1),
J,
In:
also
[CHAP. 12
IMPROPER INTEGRALS
262
2.
(a)
g{x)
and
0, and
g(x)
if
= A ^
4^
lim
or ,
then
dx and
f{x)
dx
(b)
If
A=
(c)
If
A = 00
Theorem
j
(ii)
dx converges, then
then
dx
g{x)
diverges,
a;" /(a;)
dx converges
diverges
fix)dx
if
if
A.
>
f{x)
dx converges.
dx
diverges.
Then
and
pgl
dx
x'
X 4^4
and
is finite
A^O
x^
infinite).
+ 25
'
"-^"
+l
be
X
Ax*
lim x
diverges since
Vx* +
may
(A
lim X
converges since
+ 25
Va:*
3.
f{x)
is
Let lim
1.
fix)
(i)
and
in (a)
9{x)
This test
it.
and
in (a)
a;''
1.
+l
e~*^.
fix)
J^
dx converges or
4.
C\fix)\dx
if
fix)dx
If
fix)
dx
is called
converges
fix)dx
but
absolutely convergent
\fix)\dx
diverges,
is
Theorem
2.
If
^^
\fix)\dx
converges,
then
fix)dx converges.
In words, an abso-
Example
1:
f -^^^dx
J
^^1
is
^"'^
Jo
-^n
^HL^ dx converges
X
5^2^ dx
is
^^
M^*
erges.
but
^0
^i?^ dx
I
conditionally convergent.
Any
of the tests used for integrals with non-negative integrands can be used to
CHAP.
IMPROPER INTEGRALS
12]
IMPROPER INTEGRALS
we
of the
define
on the right of
divergent.
If the limit
it is
Similarly
a^x^b,
if
=a
aSx^b,
of the interval
f{x)dx
(4) exists,
=
we
lim
call
f<x)dx
then
(A)
f{x)
we
then
a;
(,
wise
SECOND KIND
If f{x)
263
a;
=&
of the
interval
define
f{x)dx
lim
f{x)dx
-.0+ tJa
(5)
In such case the integral on the left of (5) is called convergent or divergent according
as
the limit on the right exists or does not exist.
If f{x)
we
then
f{x)dx
The
= Xo
of the interval
aSxSb,
define
lim
fix)dx
lim
f(x)dx
(6)
integral on the left of (6) converges or diverges according as the limits on the right
made
in case f{x)
may happen
zero independently.
that the limits on the right of {6) do not exist when e^ and
In such case it is possible that by choosing t^ = e^ = e in (6),
f{x)dx
^limjj'"
f{x)dx
+ C
e^
approach
i.e.
f{x)dxi
writing
(7)
the limit does exist. If the limit on the right of (7) does exist, we call this limiting
value
the Cauchy principal value of the integral on the left. See Problem 14.
C"
J
o
dx
(x
(b
- x)"
SECOND KIND
it
p<l
and diverges
if
^'o^verges if
p<l
and diverges
if
p^l.
<*''"
verges
dx
r'
a)"
of the
g 1.
p^O
the in-
CONVERGENCE TESTS
The following
the interval
IMPROPER INTEGRALS
a^x^b.
where a<Xo<b.
for
where
of the
if f{x) is
SECOND KIND
unbounded only at x = a in
unbounded at a; = 6 or at a; = a;o
f(x) is
[CHAP. 12
IMPROPER INTEGRALS
264
(a)
Comparison
1.
Let g{x)
Convergence.
Then
verges.
SO
a<x^b,
for
for
(i^f{x)^g{x)
if
a<a;gb,
~=
dx
con-
also converges.
/(a;) da;
g{x)
J,
(&)
= -!),
Divergence.
1,
also converges.
Let
g(a;)
SO
for a
<x^
b,
g{x)
dx diverges.
also diverges.
f{x)dx
a
Example:
= 4),
"^^ dx
(^
^3
for
a;
> 3. Then
since
j -^^^
= 3,
also diverges.
"^l
2.
(a)
If
a < a; g 6, and
SO for a<a;g6,
/(x)SO
and gix)
gix)^0
fix) S
fix)
dx and
(b)
If
A=
(c)
If
A = 00
in
gix)
and
(a),
and
in (a),
^
fix)
= A^
lim
lim^
if
or
>,
then
gix)
dx converges, then
fl'(a;)
dx diverges, then
fix)
fix)
dx converges.
dx diverges.
This test is related to the comparison test and is a very useful alternative to it. In
particular taking gix) = l/(x - a)" we have from known facts about the p integral the
Theorems.
(i)
Let
lim (x-a)''/(a;)
A.
Then
x~*a +
fi^)
d^ converges
fix)
dx diverges
if
< 1 and A
is finite
S 1 and A ^
(A
if
may
be
infinite).
If fix)
becomes unbounded only at the upper limit these conditions are replaced by
those in
Theorem
(i)
(ii)
fix)
lim ih -xYf{x)
Let
4.
dx converges
/(a;)da;
diverges
if
if
"^
Example
1:
Example
2:
?)<
pSl
B.
and
and
is finite
B^Q
converges since
^
Then
iB
may
^lirn^ (x
diverges since
be
- !)"'
lim
(3
infinite).
,4
- x)
_ ^y,^ ^^^ \
a;*
1 ~
";==
"2"
CHAP.
IMPROPER INTEGRALS
12]
fix)
if
Theorem
If
converges.
\fix)\dx
J^
then
dx
is called
absolutely convergent
"
"
265
If
f(x)dx
converges but
C"
"
^b
f(x)
dx
\f{x)\dx
diverges,
"^o
is
5.
\f{x)\dx
converges, then
f{x)dx converges.
J^
sin X
Since
it
\
'
^
-=
W^^\
follows that
lutely),
Xdx
X _
/
and
y/^r^
3.
^/
k^do,
IVa;-ii-|
J
'^n-
Arr
"^
'
f
^^
^^do.
a-v, p = J),
converges (abso-
\lx-Tr
Any of the tests used for integrals with non-negative integrands can
be used to test
lor absolute convergence.
IMPROPER INTEGRALS
o
IS
of the
THIRD KIND
PARAMETER. UNIFORM
Let
^()
f{x,a)dx
(5)
We
a^^a^a^, or
briefly [a^,a,].
Definition.
in
[a^,
aj
if
for each
find a
4y{a)
J^
fix, ot)dx
an
The above
<
.^()
for
all
M>
iV
and
C fix,a)dx
all
a in
[a^,
>
we can
aj
C fix,a)dx
which
is
terms.
IMPROPER INTEGRALS
266
SPECIAL TESTS
1.
for
UNIFORM CONVERGENCE
Weierstrass
M test.
(a)
g M{x)
\f{x,a)\
we can
If
of
INTEGRALS
M{x)
find a function
a^^a^a^,
[CHAP. 12
such that
X>a
M{x) dx converges,
a
f{x, a)
dx
Since
P^^ =
is
a^^a^ a^.
Example:
is
As
^"*^
"iTTT
converges,
follows that
it
values of
all real
da;
a.
2.
Dirichlet's test.
(a)
4i(x)
it
is
Suppose that
is
a;-^ 00
(b)
\f{x,a)dx < P
for
u>a
all
and
a^
f{x, a) ip{x)
^ a S a^.
dx
is
THEOREMS
Theorem
on
a^^a^ a^.
6.
f{x,a)dx
is
if
f{x, a)
dx
continuous in
is
uniformly
a^'^a'^a^.
In
particular, if a^
is
any point of
aj
^ a g a^, we
f(x,
If
ttj
is
Theorem 7.
Under
a
to a
we
a)dx
a)
dx
(9)
we can
6,
from
to obtain
I
4,{a)da
-^1
/(X,
a)dx^da
which corresponds
Theorem
to a
1)1
f{x,a)daydx
(10)
8.
If f{x, a) is continuous
dx
not oepena on
a,
converges uniformly in
a^
g a g a^,
then
to a for
if
a does
"
f
da
If a
lim f{x,
Theorem
the conditions of
can write
is easily
ur^^dx
a oa
(11)
CHAP.
IMPROPER INTEGRALS
12]
EVALUATION
of
267
DEFINITE INTEGRALS
LAPLACE TRANSFORMS
F(x)
JL{F(x)}
is
f{s)
J:{F{x)}
e-'-F{x)dx
One
is
sin
n =
'-^'^
ax
s^
+ a^
s'
+ a'
^^
sJ:{Y(x)}
Y"{x)
^ n
s>0
1,2,3,...
Y'ix)
Problems 34-36).
tions (see
cos ax
x"
*>o
e"
{12)
defined as
s'j:{Y{x)}
Y{0)
8Y{0)
Y'{0)
definitions
to
improper
multiple integrals.
Solved Problems
IMPROPER INTEGRALS
1.
dx
\/x{x
_xdx
dx
x^dx
x^ + 1
(a)
(6)
(c)
This
(d)
(e)
is
unbounded at
is infinite
a;
and x
= 1).
and integrand
is
3SxS 10).
This
is
''
lim
~
'
cos X
a;
integration
dx
X''
+ tanx
is
COSX
+ 1)
is
a;
= 2,
but this
is
-1).
bounded).
_L
= ^1 by
applying L'Hospital's
rule).
IMPROPER INTEGRALS
268
[CHAP. 12
dx
J' y/x{2 -
into
(a)
(a)
r
I
dx
-'i
yV22/ -
0<e<l,
where
- a;)
V(2
kind,
first
(b)
a proper integral.
Let
say.
x)
=1
As e->0+, we
is
equivalent to con-
X/7
-
which
y^/2y
is
first kind.
1
,
We
J.i
sideration of
r
I
VeVv^ +
dv
^0
tJv^
which
a proper integral.
is
+2
From the above we see that an improper integral of the first kind m,ay be transformed into
an improper integral of the second kind, and conversely (actually this can always be done).
IMPROPER INTEGRALS
3.
may
FIRST KIND
of the
Prove the comparison test (Page 261) for convergence of improper integrals of the
first kind.
Since
OS/(a;)
= g(x)
for
S a, we
a;
S
But by hypothesis the
hypothesis,
f(x)
fix)
^-^ = A >
44
= ^ +
and hence
(A
or
g(x) dx
>
we can
0,
gix) dx
- e) gix) S
r"
I
fix)
f
M^
\g(x)
I
A <
I
fix)
A=
A c >
and
A=
=
>,
a>
e)
gix)
(A
e)
and
p(x) da;
(1)
0.
dx exists, and so
fix) rfx
(A
/(x)dx
r"
I
fix)
so
(i),
dx converges
lim
f(x)
dx converges
r(x)dx
no
If
f(x)
Then given
any
0.
(A-e) r
If
dx
Then
is
g(x)
A -
There
g(x) dx
81,
Thus
exists,
i_fr(x)
when x'^N.
Page
7,
on Page 262.
test (a)
lim
dx
dx
lim
4.
f(x)
fix)
(i),
dx diverges
is
in general a
marked
similarity between
CHAP.
5.
IMPROPER INTEGRALS
12]
Method
(a)
For large
^'"*^
x, the
""'"^
f^ ^^^tx^ +
integrand
3a;*
Sx^
2:
verges,
Let /(.)
^"'^
+l
f(x)
3^,^^^,
sT^
converges.
(6)
Method
Method^:
Let
j-/
integrand
^ ^^-
"the
Since
= 3),
it
follows
we have
U"?
,(.)=!.
and
obtain a suitable
/%(.) d.
con-
test.
approximately x'/^/^
''""
to
is
/g =1,
lim
Hence by Theorem
l/3a;^
^^ converges.
ff{x),
|.
is
'''''''''
Then
1,
Page
1/x.
how-
f.^r/'^
1
since
It could,
1,
also diverges.
\(x) dx
and
dx also diverges.
]im
Since
3:
/(.)=-|^.
f(x)
x, the
Ml
^''''''^''
Method
>
For large
diverges,
g^.^
+ "-
3^4
jirn x'
3:
^^^
'
.(.)
Method
approximately x/3x*
is
Method
269
^-|=^i^^ =
a;
1,
the required
diverges by
integral
Theorem
1,
Note that Method 1 may (and often does) require one to obtain a suitable
inequality factor
(m this case i, or any positive constant less than 1) before the comparison test
can be applied.
Methods 2 and 3, however, do not require this.
6.
Prove that
e-^^
Jim xU--'
dx converges.
(by L'Hospital's rule or otherwise).
7.
Compare Problem
Then by Theorem
Chapter
10(a),
1,
with
A=
0,
= 2,
11.
^"^
IT^'^'ITT^ =
<^
is
a positive constant;
Theorem
"
H'^'=
l^'-'^^dx + J"L
*>y
1.
(&)
Page
-*-
262, with
""
A=
dx.
,
= l,
diverges.
(,)
fjl^^ax
The
first
Smce
Page
,
-dx
hm
262, with
cos x
x"
x^'^i^
A-0
and p
j =
0,
= 3/2.
integral converges.
Theorem
1,
[CHAP. 12
IMPROPER INTEGRALS
270
da;,
(a)
= 2/. Then
Let X
(&)
J_ ^sTi'^^r^V-
i/j
Method
1:
y^l. Then
e"" for
since
e-" dy
converges,
^dy
Ji
Ji
converges; hence
Method
Page
y^(]=
\\m
2:
262, with
A=
- f ^^f^dy.
+1
^d
becomes
^t^^
lim '(
'^'*'
9-~
Letting x
1^4^ "^^^
lim y^i
Since
2/
Since
integral converges by
0.
Theorem
1,
= 2.
and p
^+T
it
\im ye~'
^ilK
2/
J^
"
^^''^
1'
= -yin
the
first integral,
integral converges.
1,
f{x)dx
if
We
have
is
|/(a;)|
ing
|/(a;)|
i.e.
an absolutely con-
convergent.
fix)
|/(a;)|,
g
|/(a;)|da;
converges,
\f{x)\dx
IMPROPER
^a
vergent integral
for
converges,
i.e.
[f{x)
f(x)
we
\f(x)\
see that
[f(x)
f{x)
^ 2\m\.
^ aj^
\f{x)\]dx
follows that
it
Then
\f(x)\dx
\fix)\]
dx converges.
Hence by subtract-
dx converges.
10.
Prove that
Method
I
52i^
I
I
^ dx
converges.
1:
cos X
x-'
Ji
xM
Method
X
1
1
for
a;
1.
test, since
x'
(a;
converges,
i.e.
^^^
Ji
dx
converges,
j-
it
follows that
9.
x''
2:
Since
lim a;'H^^
X
a:-t-oo
= Hm
\
P^^ldx
a;-*-Qo
F^
(
-*
0,
it
follows from
Theorem
1,
Page
262, with
^^^dx
converges (absolutely).
A=
and
CHAP.
IMPROPER INTEGRALS
12]
-^^dx
Since
we need
~dx
converges.
-^^^r +
SI"
f"cosx,
as
(1)
_
-
^rf^
a'
M^
j^
and
lim
^HL
= A
r"cosx,
cosl1
cos X
^^^ =
0,
-^dx
+ j^
1
lim
(2)
(2)
2:
a;
!;
+ n?7,
sm X
if
-da;
J,
is
cosM
>
This
S1
Letting
a;
f^iilL^da;
Method
<
continuous in
is
^^^^dx converges.
/j
Method
(because 5HL
converges
271
an alternating
+ nw
^j^^ s
Since
series.
J_sinjy_
v
Also,
J,
lim
n-*
+ n^
^ ^^^
^^^
^
=
.
'"
r^^d*
+ wt
^0
Jo
1^
lim
i'
and sin^
t;
+ 2;r
sin v
r'^
Jo
Jo
iT
(%
"^^
1)^
r"*^
njT
-i.oo ./o
Thus each term of the alternating series is in absolute value less than or equal
to the preceding
term, and the reth term approaches zero as n^ =o. Hence by the
alternating series test (Page 226) the
series and thus the integral converges.
12.
Prove that
dx
converges conditionally.
As
i.e.
in
\~-^ dx
Problem
2,
\^\dx
N"^
Vhi
= (d-lV
'
*
fr
n?o
(^TflV
-I
*;
2 f
?^\dx
=oJ,r
g ^ S ..
f ^^+^''"
sin'i;
'^'^^'^^^'
is
r -^r^d.
v + nv
~=oJo
^^>
Hence
^
-
OT+TV
r"
Jo
"""
dv
it
we have
Jo
^'""^
that
diverges.
Method
11,
we must show
(n
(1)
l)n-
(2)
test.
IMPROPER INTEGRALS
13. (a)
Prove that
The integrand
converges and
unbounded at
is
^^
li^
of the
v^ + 1
^-1
a;
= -1. Then we
Determine whether
lim(^+i^'
14.
[CHAP. 12
IMPROPER INTEGRALS
272
6,
converges
,^ _^-^y
limfe-fe-)
=^
(a)
(b) in
the Cauchy
(a)
By
definition,
dx
C'
..
dx
f'-^'
dx
r'
,.
32
_,
and since the limits do not exist, the integral does not converge in the usual sense.
(6)
Since
15.
is 3/32.
,,
r'
r\
(a)
Um
^hm
dx
/-r
(X
2W3.
I)
verges by Theorem
(6)
lim x'
i_>0+
(c)
^^^^
1
^,(^3
_ 8)2/3
3(i),
dx
c'
.,
Page
lim
i(
5:
J^^^ x\x^-.tx^ A J
Since
lim
X-*.
Since
(a;
lim
V(5-)(a;-l)
Hence the
3(m) on
Page
264.
1)"^
V(5
(^
- x)(x - 1)
"*"
a^)'^'
- = -
V(5
- a;)(a; - 1)
di
I
-'s
^,
V(5-a;)(a;-l)
the
first
- a;)
^r^
ir.
integral converges.
ir/2
integral converges.
'^
2'~'^
(1
264.
lim
x^
^1
(d)
g^
1.
dx
,.
integral
con-
CHAP.
IMPROPER INTEGRALS
12]
273
Another method:
2'
2""
'
2'""
r'
1-x ~ 1-x'
(e)
16. If
lim
(r/2
-%7T-
simultaneously and
mgl
For
Se a; Ss 1.
(6)
<m< 1
using Theorem
and
<
Page
3(i),
Page
4(i),
x-^-^{l-xY-^ dx
If
m g 0, hm
<
1,
x'-^
l-m
p =
-w
^(1 - x)- =
00.
and
Hence the
Letting X
= 1/^,
"
n>0
and a
'^
of the
f ^dy
a;"-' (1
- sc)"-" =
x)'-'
using
simultaneously.
regardless of the
= 0.
^^^
""''''''^'^ ^' *
*^*" ^'^^^^-^
^* -'
-< ^^^
^-*^--'
^^''^
12.
THIRD KIND
^^'
converges conditionally.
IMPROPER INTEGRALS
lim -
=1
continuous in
""*'
1.
m>0
''^'^""^ '^^^^^"
lim
is
(i)
'""''
m>0
if
- xr-^ dx
anda = 0.
and 6
integral converges if
requirfrS'fS.Lr'''
{1
264, with
are
converges
(a)
diverges otherwise.
x-^a-xr-'dx +
264, with
'
Theorem
^
*i,
Hence
the integral converges.
j
If
real
diverges.
lim ( '^dlzL^Y" _
~
-i-\ cosx J
dx
T^^
(cos^)""
'
and n are
and n >
(a)
- xy"
^^^
(a)
converges
{J ^i^)^^^^^""'^'^^'
if
n>
and
(6)
diverges
(a)
If
wai,
and a
(J)
P"r'
converges
^'\" '"
^a=-V*'-'^f^''^
e
_ 1, the integral
a;
x^'-^e'^dx
(ij
is
continuous in
'"*'^"^' *
by Theorem
3(i),
*^
S x ^ 1.
^^'^"'^'^ ''i"'^
Page
264, with
^* -
= o-
1-rj
= 0.
Thus the
fii-st
jim
lim a;"."-"-^
a;
a;
e
Page
a;"
"-''.
Since "lir
smce
^nm x
x^-^e'^dx
'e-'^dx
262, with
**"'
for. Jo"'
?i
> 0.
1'^
^ ^^^P^"^
-%Tvt'H''Tr'"
(by L Hospital's rule or"otherwise),
this
'"^^^''^l
''^
t^
^^^t kind.
integral converges
Since
by Theorem
l(i).
= 2.
'""''
'"*''''^'
>0, and
[CHAP. 12
IMPROPER INTEGRALS
274
(6)
lin^O,
If
Page
w^O,
[Theorem
Page
264].
[Theorem
l(i),
sum
also
3(n),
Jim x-x^-'e''
converges since
(1)
Jjm x-x-'^e-'
diverges since
(1)
262].
Since the first integral in (1) diverges while the second integral converges, their
diverges, i.e. the given integral diverges if n = 0.
interesting properties of the given integral, called the
Some
in
Chapter
UNIFORM CONVERGENCE
19. (a)
Evaluate
= f
<^(a)
(c)
Explain
(o)
<f>(a)
a^a^>0.
in (a)
why
lim
ae-"''dx
e -ax
lim
Method
a>0.
for
Jl
(6)
IMPROPER INTEGRALS
of
ae-'"'dx
(&)
gamma
13.
>
to 1 for all a
lim 1
e"""
>
0.
a>0.
if
0.
using definition.
1,
a^a^>0
The
pending on
but not on
such that
a,
ae-<"dx\
r'
(1
<
ae "'dx
- e-"")
for each
if
for
e-
we can
find
N, de-
u> N.
all
<
e-"!"
>
u>-\n-1
for
N,
Method
say X
(c)
2,
Since
lim x^
Taking
Ko.
ae'""
=0
M{x)
is
uniformly convergent to
0,
the
ai
number
&
for a
Af in the first
20. If
^Wj =
I
'-'a
dx
f{x, a)
>
aj
we can choose
0,
integral
As
test.
for a
>
aj
method of
>
-^
converges,
it
0.
(6)
0.
a^^a^ a^,
is
<
|ae-"l
prove that
<^(a)
is
Then
= C
<f,(a)
h)
0(a
0(a
<f>(a
f{x,a)dx
h)
f{x, a
0(a)
where R(u,a)
R(u,a),
h)
Thus
dx
R(u, a
+ h)
and so
^)
/(ic,
/(X, a
/l)
/(, a)
{/(a;,
f{x,a)dx.
a)} da;
i?(M,
dx
|iS(M, a
^)
R{u,
a)
"u
I
/l)
0(a)
is
a^^a^ a^, we
uniformly convergent in
u> N,
R{u, a
+
,
, ^
^
<
h)\
/o
e/3,
d/ d)\
R{u,
i
I
/l)
I
^
<
+
e
>
(i)
fi(M, a)
I
0, find
/o
independ^>^
e/3
(^)
Since f{x, a)
is
continuous,
we can
\f{x,a
find 8
h)
>
fix, a)
corresponding to each
I
dx
<
e/3
for
|;i|
<
>
such that
()
CHAP.
IMPROPER INTEGRALS
12]
Using
and
(2)
Note that
a
= !,
we
(3) in {1),
0(a
+ h) -
0(a)
<
for
<
|fe|
S,
we assume
in this proof
>
for example, A
see that
275
Also note the analogy of this proof with that for infinite
sg
a,
Thus
q,^.
if
series.
21. (a)
Show
that
lim
(ft)
Since
^(a)
guarantee that
22. (a)
(b)
Prove that
ae'"' dx
by Problem
Odx
lim ae~'^^\dx.
e-"""
Thus
0.
rxdx = ^^"^2
lim
a>0
for
result follows.
(see
may
96(a)
and any
Problem
19),
Prom
M^
(6)
and
e'^^dx
EVALUATION
of
no
real value of
r.
a^a^b,
a^
+ r'
^
a^
+ r^
e'"" cos rx
e""''
converges.
DEFINITE INTEGRALS
In sin
The given
is
we have
"' ^^ ^'" rx
lim ^
e-^'cosrxdx
there
Prove that the integral in (a) converges uniformly and absolutely for
where
< a < & and any r.
(a)
(a).
19(ci).
is
cos
(b)
0.
4>{a)
lim 1
ae""') dx
( '''?.
ae-'^dx
ae-^^dx
(6)
lim
ic
~ In 2.
da;
In
^0
sm a;
da;
Letting
In cos
Jo
y dy
Tr/2-y,
In cos x
dx
Jd
Then
2/
Letting
2a;
J"
a;
In
sm 2a;
^{I
In cos
da;
dx
a;)
C'"\nf ^^^^^\dx
In 2 da;
-J
In sin
2a; da;
-| In
(1)
= v,
In sin 2a; da;
J^
Hence
(In sin
(1)
becomes
2/
In sin v
I)
- 1^
In 2
^
=
or
dv
/
7
i.J
(letting
- In
2
In sin v
2.
= TT-u
dv
In sin v
dv
>
24.
[CHAP. 12
IMPROPER INTEGRALS
276
Prove that
Let X
g- In 2.
or J
y.
= -
dx
In sin x
a;
a;
In sin X da;
~ **)
TT
TrMn 2
('^
^" s""
In sin x da;
"
a;
('^
In sin x
1" sin x
')
dx
dx
= - y In 2.
-2
+a
(b)
Show that
(d)
Prove that
^{a)
-^
r"
J^
dx
(^2
Evaluate
(c)
i)n+i
is
r'"'^
7^2^11)2
J
o
The
result follows
J^
test, since
g i.
1-3-5..-(2m-1)
cos^^^d^
^5-47^
^^T^Tf
^^ "
TT
2"
2-4-6...(2)
1
(a)
''^
>
=
/I
*'^**
r
J^
^"^
x'
+l
converges.
(6)
0(a)
(c)
From
(6),
lim
-tan
lim
= -^^
f -^^
X +a
2va
'
tan
lim
-71
'
we have
Jo
"
a^V^?+^y
Jo
^-^ g ^^
(because
8,
Page
and
j^
(a;^
+ ar
266, since
^^
j^
^tf^
is
converges).
^'
(?ny ~
,
-3-^ = "'"
n times, we
find
J."
(-l)(-2)...(-n)J^
where
r'^
Jo
Substituting
26.
Prove that
J^
a;
(x^
^^r^^.
(c).
\r2A""2A"2/-\
Letting
a^l+, we
dx
l-3-5--(2w-l)
TT
2m!
tan
l)"+'
9,
f^cos'-ff d9
22 and Theorem
7,
Page
266,
we have
;2"
find
l'3-5---(2n-l)
;r
2-4-6---(2)
From Problem
is
obtained.
CHAP.
IMPROPER INTEGRALS
12]
/"{/
e ""
cosrxdaldx
Ct
277
}
e-"' cos
= a l^*^x=0
''"-'""
27.
Provethat
J^^
By Problem
j;^"^
22 and Theorem
Jo
sec
a;
Page
7,
266,
cosrz
;3
X
28.
Since
7(6)],
it
e'"'
cos X
dx
f
a^
f^
J^
'-'
+ r^
-dr
tan-'-
5-fln( + r)
r tan
=1
^.
cos X
for a
function of a for a
fe 0,
a;
and
f\-^^l^^dx
(Theorem
6,
Page
266).
^1? dx
is
a, 1
29.
cos z
;r-dx
fl-cosa;
J^
-^i
Provethat
Integrating by parts,
'^-
Then
limit as
e -^
0+ and
-^
>
f'l-cos^j^
ax
Jo"
sm' X
da; = 2
J^
a;"
x^
IT
f
in.
lim|tan-.i-|l(^
+
,
l)|
,,
-
a;
f"
I
sin' (a:/2)
da;
-J
*'
cos X
Jo
c/O
si
shows that
a^0+,
we have
Jo
Since
letting
J.
we"have''''"*''"'"^
drXdx
to r yields
I
J^
da
e-" cos rx
+ r^
C..r
tan-'^dr
The required
C^S^
Provethat
'^^
da
a;
-'o
a^
we have
e-^'cosrajdx Idr
r"e--sinra^^
or
tan-'l
==
da;
lln*^ +
dx
ra;
rx dx
_
_
dx
s-
f'sin^tt,
-^dw
/o
**
on letting m
= a;/2, we
also
have
30.
[CHAP. 12
IMPROPER INTEGRALS
278
Prove that
ax
-r-
2i
'^
(2i?
1 /e3tx_g-3i;. \
e'-e-
1
-7
4V
2i
4^
"^
sin
+ 3
3a;
2i
sin
a;
Then
(
snr
^ax
du
5Hid^
4\2j
i J
i Jo
i Jo
4V2
MISCELLANEOUS PROBLEMS
31.
Prove that
e'^'dx
By Problem
e-'^
^7/2.
e-'
Let
converges.
the integral
6,
dx
dy and
let
lim 7m
7,
Then
^M
J,M
e-u2+v2)dxdi/
I
^M
where
12-1).
Fig. 12-1
(see Fig.
the square OACiS? of side
Since the integrand is positive, we have
%u
is
rr
e-(>.^ + ^)dxd3/
in the first
e-o^pdpd^
I
1\
jj e-''^ + ''''dxdy
circles
having radii
and
(1),
II
^0 =
e' pdpd4>
(2)
^'i-'" '^P=
^p=o
or
2^(1
32.
M^
limit as
Evaluate
e~^^ COB
Let
in {3),
7L
we And
S ^(1-e
lim 7^
7^
(3)
^)
7r/4
and 7
vV/2.
axdx.
7(,
-e -"^)
limiting pro-
cedures,
d7
e--^cosaa;da;
The
C xe-^
xe'""^
sin ax
and
dx
is
integral sign
is
justified
xe''^ dx
converges).
all
by Theorem
8,
- ^l
xe"'' sina9;|
CHAP.
IMPROPER INTEGRALS
12]
From Problem
|e-'^osaa;|
m=
lim 1(a)
Solving
33. (a)
and
e-^'
(since
i^/^.
= -|/
Prove that
j^
dx
g-^'
279
/()
subject to 7(0)
^,
^\-,.-.i.,^ ^x
we
find
/()
^.
^e-<^''\
Evaluate
(6)
f^-C-^^"^)
da;.
e-<--/*)2(l-/a,2)da;.
D
^""^^^
,-.nJ?L^w?"^*'
- 0+ and that for sufficiently
-'".f?"^
large
a;
Now
x,
by the Weierstrass
f e-^ dx
test, since
Jo
Jo
Jo
a;
(b)
From
{.
fe-^--%^
Then
34.
(a)
e-"e"dx
1
^{cosaa.}
^^
e-cosaxdx
(b)
^{cosax}
lim
e-<^-)M
-^^
= 1. r^-(^^-^-\. = :^,-,
Putting
=^,s>a;
^{e-}
J:{e'"}
f e-
,.
(6)
To determine
let
converges
-^^, s> 0.
e-^'-'^'dx
by Problem 22 with a
= s,
a.
From
part
(a),
Jl{e-'}
= -1.
s
JL{e''"}
.^{cos ax
1
Replace a by
s-ai
+ ai
s^ + a'
s^
.^ (cos
justified
ax}
^{cos ax}
sin aa;}
Then
ai.
+ a" +
'
-~^,
'
i^{sinaa;}
s^
a^
.^{sin ox}
=
17.
"
^
c,
35.
By
definition,
assuming that
Let
I7(a;)
=
=
^{F"(a;)}
r e-" Y(x) dx
Jo
lim e""" Y(M) =
Y(0)
s"
and so
Since
e-^'Y{x)
dxV
^{Y(x)}
F(0)
^{F(a;)}
s^{f/()}
Y"{x)
Y{x)
x,
= 0,
7(0)
Thus
U{0).
Y{0)]
Y'{0)
Y{x)}
<{Y(x)}
= ^{x},
- s F(0) -
j(L{Y"{x)}
F'(0)
we
J^{Y{x)}
j:{Y{x)}.
Y'(0)
= 2.
Y'(0)
J^
by methods of partial
j:{U'(x))
(o),
j:{Y"(x)
s Y(0)
M-^00
8.<:{r(!e)}
e-"Y'{x)dx
M-+00
Then by part
^'(a;).
lim
lim ]e-''Y(x)
such that
s is
s'Ul{Y(x))
-m
Jo
j:{Y"{x)}
(b)
e-"'Y'(x)dx
(6)
Y{0),
.=o
JiiY'ix)}
36.
[CHAP. 12
IMPROPER INTEGRALS
280
Then by Problem
35,
1/s^
1/s^
find
fractions.
-^^ =
and
J^{x}
.^{sina;},
it
follows that
;?
+ ^?:pY = ^{x +
sin x},
smx).
Y(x)
a;
sin
Another method:
If
j:{F(x)}
By Problem
f{s),
78,
we
.^-' {/()
S'W)
-C"' {/()}
Then from
^"Mfl'(s)}-
f(s)
.C"' {F{x)}.
(i),
Inverse Laplace transforms can be read from the Table on Page 267.
Supplementary Problems
IMPROPER INTEGRALS
37.
of the
FIRST KIND
,,,
fr+yd*
J
f"
(d.)
xdx
,,
f"
dx
-^^
2
sin
^^dx
r"
,,
(ff)
!^
j_^(x'
,.,
^1
a;V3a;
+2
(o) conv.,
"^^
(6) div.,
(c)
conv.,
^'"
a;
ax
'^
"'^
(d) conv.,
SC
ln^
X dx
sin'
Ans.
+ x + iy'^
(e)
conv.,
(/) div.,
(g) conv.,
(h) div.,
(t)
conv.
'
CHAP.
IMPROPER INTEGRALS
12]
f -^p-
38.
Prove that
39.
Ans.
(a) conv.,
?L_
:=.
(a)
(6) conv.,
281
ifh^\n\
(b)
e-^r.
(1
e')
dx,
(c)
(c) div.
40.
f ^i'^
+
(a)
x'
Jo
J_^~"'cos6i;da;,
^*^
r_l^rfa;;
^ Vx' + l
(c)
^=0
/
cos
>
^^^
a;
^'^^^
41.
IMPROPER INTEGRALS
() abs. conv.,
(6)
and
..
Ans.
-^
dx
(a) conv.,
(/)
(b) div.,
TTib^Jr^
X
Ams.
a;
^^.
v^M^
(d) div.,
(e)
abs. conv.
on Page 262.
(c)
^-^Tr^'^^
i 3in
J^
. dx
(d) conv.,
ii>
(.)
conv.,
f ^Jl^
(/)
(^)X'S
Jo
Vln(l/a;)
(c) div.,
fl^
Lv.,
ax,
(^) div.,
\kl
(/.)
<
div.,
(i)
conv.,
cos(l)..,
'"""^"' "^'"^
Prove that
(6)
TYso:^
IMPROPER INTEGRALS
(a) conv.,
cond. conv.,
sin
i - x
C^)
cos
(c)
div.
V^
(a) conv.,
Provp
tVint
^rovethat
^^
conv.
(c)
-V/^M^
a > 2, div. if
(6)
conv.
if
Jo
sb (ax)
r 5-h^^dx
,
J^
converges
if
<
^''^^
(b)
As.
X'^cos(i)...
(e)
{a)C e~'lnxdx,
(6) div.,
(g)
V^M^+1)
C"
'
-^
(3
'
dx
2 sin x)
"
Vsinh
(ax)
'
2.
|a|
<
;.
and diverges
if
|a|
&
;r.
l^'^^'
(*)
conv.
THIRD KIND
of the
Ans.
X |cos(A)..,
(.)
O')
*'' ''"""''''
49.
cond. conv.,
(c)
48
abs. conv.,
(6)
'"'Iw^^'^
'^o
46.
sin
SECOND KIND
of the
^"^X^^Twf^
45.
a:
42.
44.
J"
'*"'
H^ih^
f"
((i)
'
dx-
lii^^^-
^--
()
-nd.
conv.,
(6)
abs. conv.
'
UNIFORM CONVERGENCE
50.
[CHAP. 12
IMPROPER INTEGRALS
2g2
(a)
Prove that
(6)
Prove that
5L Let
ie
^(a)
4.{a)
= jy(x,a)dx,
lim
rF(x,a)dx #
52.
Work Problem
53.
If F(x) is
51
where
(c)
all a.
F{x,c.)
Find
- <
in
F(a;).
Theorem
and
(6)
54.
Prove
55.
56.
Prove that
57.
Prove that
(a)
<
a;
0(a)
is
not continuous at .
(a).
and
yF{\)d\
r"
so
e-"^
'-^
e'"' F{x) dx
dx
g 0,
(6)
,(.)
59.
of
63.
bx
dx
In (b/a),
dx
tan-'(6/r)
j^^-L^^dx
f ?""Jd
8,
Page
a,b
Justify
all
>
X
'
61
Theorem
DEFINITE INTEGRALS
=|-tan-a,
58.
60.
S 0.
EVALUATION
= 0,
converges, then
(a)
r/2.
on Page 266.
jf^(^)^^
r"sinx^^ ^|.
Jo
Theorem
test for
if
that 0(.)
(6)
lim V{x, y)
Show
(c)
a^xe'"'.
prove that
(a)
all a.
Ans.
lim ^(a).
<^xe-'"'.
lim F{x,a)dx.
f"
F(x,a)
if
is
f^''^
j^
continuous for
is
i>{a)
IMPROPER INTEGRALS
of
e~
^
"
X cscrx
>
a,b,r
tan-Ma/*-).
|H
/cosaa;
cos ba;\
i,Y "'
^'
X"
'
(b)
Use
(a) to
prove that
[The results of
{F(0)-F(>)}
66.
Given
C
-'''
(6)
dx
-cos
6a;
^^
,^^b
integral,
and Problem 60 are special cases of Frullanis
In^-Y
e""'"
cos ax
which
is
IV^iT^,
/"
I
>
,,
a;2pg-ax2rfa;
13
=:
2*2'2'"'
1, 2, 3, ...
(2p
- 1)
2
V^
"''
2 ''" +
J^
ax
a.
'
CHAP.
IMPROPER INTEGRALS
12]
tan
Prove that
69.
'
(x/a)
tan-' (x/b)
- JiL.
~ T~r'
'^^
f"
J_^(x' + x + lY
_
-
^^
tt
283
/6\
"^^^^^
^\a)
rir- *
tt
n u,
[Hmt:
Use Problem
a>0,b>0.
38.]
MISCELLANEOUS PROBLEMS
jlM^rrf^ eonverges.
70.
Prove that
71.
Prove that
i
J.
x'sin^^
converges.
[Hint: Consider
n U X
, n^
A
,
^7]-
j^
X
bill
J;
that
J^
72.
Prove that
73.
(a)
Use
sin^
cc^
J,_^
{n;ry sin^
g
f/j^f^dx = . In +
show that
C" In sin d# = -Z in
(1
(a) to
),
Prove that
75.
Evaluate
Ans.
76.
(a)
(a) If
(a)
(6)
(6)
78.
79.
^{F(a,)}
6^
/(s),
Prove that
>
(c)
^{e'F(a;)}
prove that
Ans.
^{x' Fix)}
^'
(b)
tan-
A\
- a),
/(s
>
0.
(6)
=
s
(-!)"/'"'(),
>
^-.(^(,)},
(6)
Y"{x)
Y'(x)
(c)
Y"{x)
2 Y'ix)
Ans.
80.
x)/x}.
.<:{(sin
a;}.
^-{/(s)
(c)
s>\a\
prove that
o
- ay +
ax},
(6) .<;{cosh
(b)
^{Fix)}^f(s).
If
(a)
^{1/v^},
V^s,s>Q
(s
77.
f^iHl^rf^ ^ E.
Ams.
0.
2.
i/o
74.
a;
j;"^^^||_ diverges.
Prove that
(6)
(a)
Yix)
2 Y(x)
6e--
= 2,
= 4
Y(0)
- 3e-^
It
(6)'
F(,)
= -3
K(0) = 0,
F'(0)
F'(0)
Y{x)'
^
is
-=
F'(0)
A- 2e' -
Ix'
2-^
- X,
x
stion.lly conllnuoas
(e)
[C)
in
Y(r^
nx)
-x/
i -I
e '(sm
j_
a;
10
cos x)
il
.h.r. H-,
81.
[CHAP. 12
IMPROPER INTEGRALS
284
and
m=Jl{F{x)}
If
g(s)
H{x)
is
Write
Hint:
F(u) G(x
where
J:{H{x)}
- m) du
F
=
/(s) ff(s)
f(s) g(s)
"'''^^^^^^}
{/f^'^'^^^^ '^^jlX
and then
e-(+f)jr(ti)G(v)dwd'y
let
+v
==
^0
82.
Find
(c)
Ans.
83.
=C-{(^}.
(a)
i(sinc.-a,cos:.),
(a)
Let
/(x), i,(x)
and
g'(x)
Y"{x)
Solve
(6)
Y{x)
Y(x)
-+
Y(.)
{'
f[x) g{x)
dx
f{x)
dx
is
- u) du.
- m) du,
a;
a and
(b)
integral on the
Prove that the integral on the right, and hence the
g(x),
that under the given conditions on f{x) and
test.
integral
called Abel's
is
For
For
(6), first
(a),
lim
consider
prove that
if
f(x) g(x)
\h{x)\<H
81.]
+ xV6
0.
S'(^)h{x)dx.
Prove that
[Hint:
lim g(x)
(a)
Use Problem
[Hint:
y(x)
(c)
finite interval
all
= y'(0) = 0.
7(0)
Yiu) sin (x
bounded for
= ~j
R^x),
i?(u) sin (x
be continuous in every
g{=c)
dx
(a constant),
then
'()
left, is
convergent.
converges and
result
sometimes
h'{x)
Hx) dx
is
The
H{g(.a)
- g{b)};
and then
J^
lot 6
A->
-> ool
let
84.
Use Problem 83
85.
(.)
to prove that
Given that
(a)
^ ^dx
and
^"sinx^d. = j;"cos.^dx = |
evaluate
sin (x^
Explain
Ans. 7r/4
j^"
^1
sino.^ dx,
^^^
j,
^'^^^^^"'^
>
1,
converge.
^.o
I
(6)
(6)
why
+ y^) dx dy
(a).
chapter 13
Gamma
GAMMA FUNCTION
The gamma function denoted by
which
is
w>0
convergent for
(see
r(n) is defined
Problem
gamma
r(n
Chapter
18,
function
+ l)
by
12).
is
nT{n)
^2)
For
Examples:
r(2)
l!
w!
r(6)
i,
5!
^(5)
r(3)
120,
mTtrirv
''"
takinf
TABLE
of
for example.
VALUES
n
and
.4!
~ 2! ~
^'^
v^
The process
GRAPH
(^)
Tin)
^^""^
7,
(5)
4) that
r(i)
See Problem
1,2,3, ...
It
+ l) =
of the
is
^^^
(i)
as a solution.
+ ^)
(5)
GAMMA FUNCTION
T{n)
1.00
1.0000
1.10
0.9514
1.20
0.9182
f(m)
5
1
1.30
0.8975
1.40
0.8873
1.50
0.8862
1.60
0.8935
1.70
0.9086
1.80
0.9814
Fig. 13-1
VV_-^
J
-3 -2
-1
2
3
1.90
0.9618
2.00
1.0000
A:
285
By
5.i
[CHAP. 13
2gg
ASYMPTOTIC FORMULA
If
is
parent.
for r(n)
0<e<l
+ 1) = V2^n-e--e'"''-+''
which
For most practical purposes the last factor,
omitted. If n is an integer, we can write
n!
is
(^)
"is
In particular
if
= i,
a;
r(i)
is
called Stirlmg's
sometimes
GAMMA FUNCTION
the
0<a;<l
= -r^^~
smfljTT
T{x)T(l-x)
n,
^/2^n''e^''
where ~ means
(^)
can be
()
= V^.
22^ - 1 T{x) T{x
This
is called
(2,r-"'TM
r(x)r(.+l)r(.+|)...r(+5^) =
The
gamma
an
infinite
W:.
V{x
where
n(a;,
A;)
is
is
an
lim
iii^(a,
llllliiJ^
fcx
The constant
function.
limn{x,k)
is
{12)
r(a;
This
=
-
l^
+ \)
()
m = 2.
This
(9)
+ ^) = V^ r{2x)
the gamma function.
+ l)
V2^^a;"e
f
HI
+ Y2^ +
the
called Stirlvng's asymptotic series for
20).
Problem
(see
asymptotic series
is
r'(l)
Ce-''\nxdx
-y
139
288^
51840a;3
gamma
where
1
"^
'
,j^x
"
'
'
function.
is
The
series in braces
Euler's constant.
{H)
T{x)
The
BETA FUNCTION
denoted by B(to, n)
B(m,n)
which
is
convergent for
m>0, n>0.
is
defined by
C x^-'{l-xY-^dx
See Problem 16, Chapter 12.
(ifi)
CHAP.
13]
is
gamma
ipiM
Bim,n)
resuJt^Tre
Qi"Ti2m-I
valid
form>0
287
and
n>0
2n-l o
/3
Problems
[see
rr^^^ =
^/i
(10)
^
i
and
ti
(13)]
functions.
T(m)r(n)
'
2r(m + )
Two
useful
(18)
and
-^
r(p)r(l-p)
gamma
o<,,<i
(19)
DIRICHLET INTEGRALS
If V denotes the closed region
-j =
if all
(~\ + ( A\
pqr
r(l+f + f + ?)
Solved Problems
The
1.
GAMMA FUNCTION
Prove:
(a)
T{n
+ 1) =
x-e-'dx
1, 2, 3,
r(2)
In general,
=
.
lr(l)
r(n
lim
in
1,
+ l) =
(6)
lim
n^
{-^"-" +
e-'do;
Put n
n>0;
nT{n),
1)
r{3)
2r(2)
r(n
n!
if
+ 1) =
r(n).
is
nl,
n = 1,2,3
x-e-'dx
^"-e-d.|
e-'d^
T{n
nv(n)
Um (l~e-)
ifn>0
=
Then
= 2-1 =
2!,
r(4)
a positive integer.
3r(3)
3-2!
3!
[CHAP. 13
2g8
2.
"2r{3)
F(I)
^"^
r(5.5)
_
~
6r(|)
(d)
5r(|)
30
i-ir(i)
3.
r(4)
u_
2!(1.5)(0.5)r(0.5)
_
"
r(3) r(2.5)
_
-
r(i)
2-2
2-2!
|r(|)
_
-
r(|)
,,,
(")
5^4-3j^
_5]_ _
_
~
r(6)
/.^
^'
315
(4.5)(3.5)(2.5)(1.5)(0.5) r(0.5)
6(|)(|-)r()
4
3
5r(f)
3.
(a).
r"xe-'d*
(6)
Let
a'e-'^dx.
r(4)
/-"/
2a;
Prove that
r(4)
5.
A/
''"''''
r(7)
6!
= -i^ =
2^
Letting x
r"e-"^dM
= u^
2(^)
this integral
= V^
becomes
^/ye-'^ dy.
r"
3-4.2^^
Letting
j/'
x,
V"
(,)
'^'
r"(e'^)'"""''dz
becomes
Let
=0'"
u.
Then x
r^iZdM
Jo
^m
Letting
g-ax
ax'^y, the
".
When
a;
dx
We
(a)
r w-'/^e-dM
r(l/2)
rl
1,
(4 In 3)z'
0;
r(i/2)
when
a;
V^
= 0, u =
The
integral
v^
^''
-Ju
integral becomes
r{"T-Hi
Evaluate
Let
-'""'"' d.
if"
- In X =
becomes
7.
45
V^.
x-^'^e-'dx.
j/.
1(0""^
4.
3!
^^^
.;""
'"*
by
use the generalization to negative values defined
r(M)
i^'"""'
^^^^^^
-
CHAP.
13]
(a)
Letting:
(6)
Letting n
=
=
-|,
r(-l/2)
-3/2,
= -2V7.
^iig^
r(-3/2)
_ -2\^ _ 4Vi
-3/2 ~
3
^(-1/2)
-3/2
Then
Prove that
8.
r(-5/2)
J^'
^^t^
where ^
(~^^'
fr
Page
177.
''"'"'*''"*'^'''''''"'^
become?"^
9.
particle is attracted
forTtirre'ch O.
^^
where
is
Lw
the
^^* '
"^
toward a
"''
''* *'''
by N^it^'s
'"
'"'
8,
--'"'"' dy.
fixed point
"
''
is
+ l)y =
(m
u,
m>-l.
*'^ ^^^^^^^
^-^ -'
^^^^^^
^^
=.>
and
let
^'^ thTtlm:
be the origin.
Then
d^ _
k
- ^
^^dt'
mass of the
'"
Chapter
50,
"^'"^
'
= "JV^-
^=^
..".(lnx)"d..
289
*^
particle
and
>
A;
a constant of proportionality.
is
(1)
~=~
Then
d'
- ~x
x = o, we
i^
mv^
2
at
_
=
-
find
'^^'~
'^'"^
lit"
t%".?l:it^:^jL'='ir"
taken for
the particle to go from
!'"'=^
=a
to
^-
"^^dx
'^
cc
A In
^
"
^----^
=
rft
dt
^
dt
= -klnx +
'^
^ ^,M,.
dJ ^"^ (') becomes
:*i
(;g)
Then
a.
*i.
da;
= -A/^>/in^
-Vi>f^
-.* ---ses.
We
(^)
given by
is
dx
Letting In a/x
The
10.
=u
or
a;
ae-",
this
becomes
BETA FUNCTION
Prove that
(a)
B(^,n)
(b)
(a)
Ix'^-Hl-xr-^ax =
(fo)
~ l~y, we
sin^ $,
B(m, n)
sin^-i
cos^-'
^ dd.
"
have
j;'(l-,)->,-.rf,
= f^r-il~yr-dy =
we have
^x-Hl-xr-^dx =
(sin^*r-.(cos^.)-.2sin.cos.d.
J^
Bin,m)
11.
[CHAP.
290
= ^^+n)
B{m,n)
Prove that
Letting z
= a:S we
Similarly,
r(w)
^''>^-
z-'e"dz =
V(m)
have
13
J^
Then
= 4(^J%^-'e-^dx)(^J\--'e-^d2y'^
r(m)r(n)
a;
p cos ^,
p sin
</>,
-co
p2<+n) -
J,7r/2
<f>
4(
"'
e-p2 cos'"
dp
sin'"-'
d^,
^p =
p"'+'"-'e-'''dpY
cos""-'<;6
d0
sin'""'
cos"""'
sin'"''0d0J
r{w +
w) B(to,
m)
X17-/2
= r(m + M)B(m, )
follows.
Chap.
The above argument can be made rigorous by using a limiting procedure as in Prob. 31,
12.
X'
({,)
x*(l-xydx
^"^'^
-'o
\/2-x
Letting x
4^2 j]
r(5)r(4)
'^^^^
B(5,4)
= 2v,
= 4V2 _
-f^dv
Letting
i/'
or
a'a;
that
sin^-"-^ 9 cos^"-' 6
3/
,-
v^(l-v)-^''dv
a'j\^'m-xr'^dx =
Show
/'
y*y/aF^dy.
13.
4! 3!
-^
r-
4V2r(3)r(l/2)
;(4)
llMp)
= 2^1^+"^^
m.>0.
aB(5/2,3/2)
de
.-.
4A/2B(3,i)
14.
Evaluate
(a)
Let
(a)
r'\m^9d9,
2m-l =
6,
2n-l =
0,
(b)
i.e.
w = 7/2,
dS,
sin* 5 cos^
m=
l/2,
^^
in
(c)
Problem
^'^.^.
fcos^^d^.
13.
g|
64V2
1^
12.
CHAP.
13]
2m- 1 =
Letting
(6)
4,
2n
291
^(5/2) r(3)
5,
The given
(c)
integral
Thus
2m -
letting
_8_
315'
2r(ll/2)
cos* e de.
0, 2ri
-1 =
2 r(l/2) r(5/2)
is
2r(3)
15.
Prove
sm^ e de
J
'^o
=^
cos^edd
-'o
tive integer,
'\':|;^-^^-
(6)
From Problem
13 with
^^
if
2m- 1 =
p,
an odd positive
is
2n- 1 =
If
= 2r+l,
2r[i(p
^^1
+ l)r(|) ^
2r(r+|)
sin^
From Problem
sin3.(l-sin^,)rf,
The method
X m^^
Letting
~o=y
J^
Evaluate
Lety* =
with p
The
_
"
-as
i^^ | = |
2.4.6-..2r
2>4-6--.2r
(2r
seen by letting
(c)
J^
^in^ ^
[compare Problem
_ r
3j3,^,
1)
./2
- ^.
d^.
14()].
^.^,^
l-3'5
M'3-6-7
^
15"
_^__2i4
1*3
sin e de
18.
integral equals
J-.Tr/2
^^- ^^"^"^
+ 2)]
sin3^cos^^d^,
(6)J^
15 the integral equals
cos- d..
j^
cos^cZ.,
^
The
^^
integer.
2r(2r-2)...2
r(r-l)...l-V^
2(r+|)(r-i)...^VF
ede
(6)
^^
(a)
tt
8'
"^^"^
(6)
1)
Zir
we have
0,
X17/2
p = 2r,
If
(a)
_
~
357r
s;
2-4.6-82y
eT'
^'
= g-^, nuw
how that
mat r(p)r(l-p)
=
nP;r(l-p)=-j-!^
where 0<p<l.
''
"""r^'
y-'H-yy^dy = B(p,i-p) =
r(p)r(l-p)
C J^
x.
Then the
integral becomes
C ^Lllax
+
4 J
1.
by letting
y^
tan
a;
e.
'^
_ ^V2
'^
4 sin
(,r/4)
4~
"'^
P^oWem
17
.2
19.
[CHAP. 13
292
Show
that
x^
Letting x^
or
%y
IGtt
dx
9\/3'
the integral becomes
2%/'^,
(\,/'^-i/W^)'iy-"'dy
lj^y-"'(-i--yy"dy
|b(|,|)
8 r(f ) r(|)
3
Show
= V2^^"e-
We
1)
now
to
"
maximum
= n + y.
for x
Then
'
= w,
(1)
as
dx
(1)
is easily
shown by elementary
becomes
In (1
= y/n. Then
= Vnv, we
on letting y
r(K
It is of interest that
1)
from
(4)
+ x)
is
=
^X
dx
(2)
When n
9V3
the analysis
be made rigorous by
them.
with x
riiW3
approximately.
X" e
has a relative
This leads us to the substitution x
Up
In
T(n
have
calculus.
FORMULA
STIRLING'S
20.
fr(i)r(t)
r(2)
^^
~"
2"
(3)
find
is
w" e""
we can
Vn
e-''2
I
dv
V^^n" e"*
Page
(5)
286.
See Problem
DIRICHLET INTEGRALS
2L Evaluate
= jjJx-'-'yi'-'zy-'dxdydz
coordinate planes.
Let
x^
= u,
y^
= v,
z^
= w. Then
rrr
~
JJ J ^
"'
du dv dw
2Vtt2\/v2Vw
1 jj
n(a/"-iv<p/2)-i ^(v/2)-icjMdvdw
(i)
where
is
Fig. 13-2
= 1 ~M
74.
CHAP.
13]
Letting d
(1
- m)*, we
have
/*'"""
(1
- M)<fi+r)
r(ff/2)r(y/2
^[(^
so that {2) becomes
lW2)r(y/2 +
+
f
^!f .'^'l^y^
4rr[(^
+ y)/2 +
22.
l)
r[(a
l]
(y/2) r(y/2)
^[^^ + ^'^'^ + ^]
+ ;8 + y)/2 + l]
^^'^''^
^>
+ y^ + ,' =
^^
x^
Show
where
Let X*
= y. Then
Page
a'-a^-a'
2-2 -2
287,
let6
= c = a,
24.
Then
J^
l]
is
^,
-^
^^^ ^^^^.^^
first
.^
_
^
"gil"
v^
{r(l/4)}'
r(l/4) r(3/4)
S
-yv -3.
4,ra
i^i^^))'
17 with p
x^y^z\
octant bounded
'
= 1/4,
sink's; da;,
= lB(p + |,l) =
=/
1 r(l/4) r(3/2)
r(7/4)
r(l/4) r(3/4)
+ ^ + y)/2 +
- /jfl
= r = 2^ ana
p=g
f
and a~
1
r(l
r 2/-"Ml-2/)"*d2,
J ^
From Problem
>
fvT^^dx
that
8r[(a
(20),
r(a/2)r(ff/2)r(y/2)
-
+ y2 + ^2^
MISCELLANEOUS PROBLEMS
23.
+ 1).
r(y/2
+ ^)/2
+ l)
+ 1]
f'
by the sphere
293
= .^2
22''-ir(p)r(p
+ i) =
(P+i)V^
2r(p+l)
v/^r{2p)
T>,.
Then tv,
the
[CHAP. 13
294
Letting 2x
= u, we
find
sin^"
(2 sin
dx
udu
X cos
a;)^"
u du
sin'"
2'"
sin'"
x cos'" x dx
J^
+ i)}'
r(2p + l)
22P-1 {r(p
=
Then
since I
2'"-'B(p
+f
+ i)
J,
"
2pr(p)
Show
that
2pr(2p)
follows.
"'^
25.
{r(l/4)}^
d4,
Consider
.^/.
r(i)V^
^^
-i=
cos d
iB(i,i)
-Yf^
{r(i)}'
^^
as in Problem 23.
^"^^
"
'
VW7
Jo
Vl-2sin'e/2
^0
Vcos'/2-sin'9/2
Jo
^""""^
^'"^'^^
V l-J-sin'0
the result follows.
26.
C"
Prove that
We
cos
J^
^
x"
have
"
iC
^ ,n
"<P<-^n
"^^^'^ =
= i;^
r(p)
f
Jo
2 r(p) cos
M"-'e-"dM.
(p7r/2)
.r-
'
Then
Jo
cos X ^^
'
_J^
r(P)
-^o
-'o
^ f^^^du
r(p)Jo
+ M"
Problem
where we have reversed the order of integration and used
17
Letting m* = -w in the last integral, we have by Problem
r"
Jo
M"
r+l?'^'*
Substitution of
27.
Evaluate
Letting
Problem
26.
x^
-1
~
(2) in (1)
r" f'"""^' J
2j ir^"''
_
~
22,
Chapter
ii-
2sin(p
+ lW2
12.
"
(2)
2cosp^/2
cos x^ dx.
= y,
X ^^ ^^
^ ivT/2
,
[see Problem 68(a)] are called Fresnel
This integral and the corresponding one for the sine
integrals.
CHAP.
13]
295
Supplementary Problems
The
GAMMA FUNCTION
28.
Evaluate
Ans.
i^
^-^
(a)
(a) SO,
Evaluate
(b) 16/105,
(c)
e-'^dx,
(6)
Ans. (a)jr(i),
(6)^,
3t
Show
32;
Prove that
33.
Evaluate
Ans.
34.
that
(a);
(a) 24,
Evaluate
(6)
(a)
35.
Prove that
36.
Prove that
v'Je-v^da;,
J^
(c)
(c)
-3/128,
r(-7/2),
*^
is
Ans.
r(-m +
38.
{g) Find
Aws.
(a)
(a) 1/60,
(6)
is
Problem
49,
(6)B(3/2,2),
a;^(l-a;)'(far,
j^
xdx
e"- In
j^
BETA FUNCTION
W2,
(c)
41.
Prove that
(6)
J^
(16V^)/105,
a negative number
Chapter
Vd -
i)/a: drr,
(b)
-3r(2/3)
(-1)" 2" V7
l-3>5-(2m-l)
1)
(it is
Ans.
(c)
(a)
1/105,
(6)4/15,
(c)
(c) 3,r
Prove that
(a)
(a)
12;r,
cos'ede.
Ans.
(a)
3^/256,
cos=9sin^dff.
Ans.
(a)
16/15,
Vtan
- x'
(6)
:r
(6)
de
V3a;
(r(l/4)r
^^
sin'edtf,
45.
Ans.
*^o
J-.T/2
'^^
(6)
{h)
J^
s-/V2.
^,
(6)
= 0.577215.
11).
B (1/3, 2/3).
(c)
tt''M4-)''^dt..
f"
v'hTaT^) dx.
X*
(c)
y'e-"' dy.
{a)
a positive integer,
Evaluate (a)B(3,5),
(6)-^
0-
m = 0, 1, 2,
The
24
| r(|)
r(-l/3).
(6)
>
(xlnxYdx,
(6)
(c)
'*'*'
where
r'(l)
(a)
s>0
(}^^
J^
lim T(x)
Prove that
Ans.
5v^
^"
37.
x^e-^^da;.
(c)
^(4/5)
-xR,
(Inaj^da;,
if to is
(c)
aj'e-'-dx,
(6)
dt
r(n)
f^-'^
f x^e-dx,
(a)
EBm^
(6)
C t^M.
= _^^
(6)
5,r/8
(6)8/105
2,r/V^
46.
Prove that
f_^J^, ^^
^^^^
47.
Prove that
jje^'+iy^''
[Hint:
48.
[CHAP. 13
296
Problem
of
31,
Chapter
--
Problem
>
' ^
"
46.]
DIRICHLET INTEGRALS
in the
49.
50.
^'"^^
yfxy-
52.
53.
54.
= 0,
if
the density
is
x-
i*
constant of proportionality
Ans. 4^/35
1.
first
+ u"'^ + z^'^ =
+ y- + z'' =
a",
1.
where
m>0,
is
given by
r(3/m)
Show that
is
1,
+ V + e%
(a^
+ y'" + z^'^ =
8 {T{l/m)y
3to2
x^'^
$+ |^+ f^
Ans.
its center.
Show
+y =
'^/24
xy plane bounded by
x-
._._-_
-
given by
,,
>
a-,
where
m > 0,
3 r(2/m) r(3/m) ^
4 r(l/m) r(4/i)
'^
+ y- + z- =
MISCELLANEOUS PROBLEMS
55.
j'
Prove that
Let
[Hint:
56.
Evaluate
57.
Show
58.
Prove that
- a)' (6 - xY dx -
x-a =
f
(a)
^"
Let y
a!/(l
iXllSp^'^''
r,
4-
[Hint:
Let
Prove
61. D
ti,=f
that
[Hint:
62.
f 'a;"'-'(l -')""'
that
Prove *i,
J^
^^^
^.
^)m+.
x-{r+ \)y/(r +
p"
J^
Let x
Prove that
sin' 9
_
-
Ans. (a).,
W ^^^^
where w,i>0.
in
PIT
TT
2 '^T'
B(m,TO)
^~(i
6 cos'
do
er^"
+ r)'"+''
where
to,
n and
_
"
B(m,w)
wherem,n>0.
2a'*6'
- jT+f'a'
^
+ 1)
j/).]
a;).]
,
tan's d9
60.
1,
(l>)
X"^'^
-.
- )-^'^' B(p +
__.
B{m,n)
(6
- a)j/.]
(6
imOL
that
[Hint:
(x
>
o.
CHAP.
63.
13]
m = 2, 3, 4,
m
[Hint:
Use
297
2""'
by
a;
1,
XIT/S
[Hint:
65.
66.
67.
Square the
In 2
using Problem
result in Prob. 63
...
63.
v(^^^^^
^
x"
a;"
m -
(^-)"""""
.
(S),
Page
Problem 65 and
result in
dx
and p
dx,
-^,
2 T{p)
sm
,,.
,
{pTr/2)
m^
let
X
1
J*
70.
0<p<l.
'^
= 1.
(6)
Jo
71.
Let
x cos
Jr(x)
x" dx.
Ans.
(a)
A\/W2,
(h)
(1)"
If
>
0,
>
cot p:r,
0<p<l.
16
(a;/2)''"'"^''
^^ n\T{n + p +
l)
'
Ptt
^^.
a;
= ^
3\/3r(l/3)
-ir^ CSC
73.
dx
-r
3.
("-^dx
+1
Show that
and 4ac
>
(c)
of
sin x,
(6)
75.
{i5)
[Hint:
Expand
Obtain the
e"*'"*^'
resiilt (15)
Page
xY' + xy' +
J-in(x)
(a;^-p*)j/
^2/vx cos x,
(c)
0.
the
all p.
v/4ac
Obtain
(16),
6^ prove that
r
74.
286.]
>.]
*^o
/v.p-1
as a definite integral.]
= ^ In {2w).
In T(x) da;
J
(6)
v/2
left
Prove that
(a)
da;
Prove that
[Hint:
a;
Prove that
[Hint:
In sin
in a
on Page 286.
power
series
6*
20.
>.]
chapter 14
Fourier Series
PERIODIC FUNCTIONS
function f{x)
is
Example
The function
1:
...
all
equal sin
"^.tt
is
The period
of sin
Example 3
The period
of tan x
Example 4
is
for
a positive integer,
is
all x,
called the
Example 2
if
is
2jr, Air,
However,
x.
(a;
+ 6jr),
is 27rln.
ir.
number
as period.
in the
m)
/(x)
rz
(b)
and
(c)
below.
rr
(c)
(b)
Fig. 14-1
FOURIER SERIES
be defined in the interval (-L, L) and outside of this interval by f{x + 2L) - f{x),
The Fourier series or Fourier expansion corf{x) has the period 2L.
responding to f{x) is given by
Let
i.e.
f(x)
assume that
ao
-^
-f,
'
coefficients an
and
\
n-rrX
& are
f{x) cos
-J
dx
w=
bn
If
/(a;)
n-irX
r"^"" L
0,1,2,
..
(^)
nirX
J-
f{x) sin
da;
has the period 2L, the coefficients a and b can be determined equivalently
from
n-nx
fix) cos
/(a;)
-^ aa;
(3)
where
c is
any
real
number.
^J^
sin
n-n-x J
-^ dx
298
(3)
becomes
(2).
CHAP.
FOURIER SERIES
14]
To determine
see that
2LJ~
co
we
Oo in (1),
r*^
is
the
(2)
or
with n =
(3)
0.
f{x)dx.
j^J
f^^^^^> which
use
299
mean
(i)
is
we
(2)
equal to
or
(;8)
{3)
The
DIRICHLET CONDITIONS
Suppose that
(1)
f{x) is defined
(2)
f{x) is
(3)
f{x)
and
Then the
/'()
+ 0) and
+ e) and
f{x
lim /(x
e*0 +
number
of points in (-L, L)
2L
(a)
f{x)
,,.
/(aj
if
is
or
(;2)
{3)
converges to
a point of continuity
+ O) +/(a;-0)
^-^^
(0)
Here
finite
.^
if
a;
IS
,.
a pomt of discontmuity
- 0) are the right and left hand limits of f{x) at x and represent
lim f{x-i) respectively. For a proof see Problems 18-23.
f{x
-+0 +
The conditions
(1), (2) and (3) imposed on f{x) are sufficient but not necessary, and
are generally satisfied in practice. There are at present no known necessary and sufficient
conditions for convergence of Fourier series. It is of interest that continuity of
f{x) does
not alone insure convergence of a Fourier series.
ODD
and
EVEN FUNCTIONS
A function f{x)
are odd functions.
A function f{x)
are even functions.
is
is
called
odd
if
called even if
f{-x)
f(-x)
-f{x).
f{x).
2x<'-4a;2
+ 5,
since,
cos
The functions portrayed graphically in Figures 14-l(a) and 14-1(6) are odd
respectively, but that of Fig. 14-l(c) is neither odd nor even.
a;,
tan
e^
3a;
+ e-^
and even
In the Fourier series corresponding to an odd function, only sine terms can be
In the Fourier series corresponding to an even function, only cosine terms (and
possibly a constant which we shall consider a cosine term) can be present.
present.
or
COSINE SERIES
FOURIER SERIES
300
function
is specified
interval,
namely ( L,
an
[CHAP. 14
0,
^n
= Y
fi^) s^^
"T~
'^^
(i)
bn
0,
an
= Y
PARSEVAL'S IDENTITY
cos-^dx
f{x)
states that
lj\f{x)rdx
L
if a
and
+ t(al+M)
and
(5)
if f{x) satisfies
the Dirichlet
conditions.
DIFFERENTIATION
Differentiation
Theorem:
The Fourier
COMPLEX NOTATION
Using Euler's
L g g L and
for
= V^
(see
f{x)
dx
x,
is
a;
FOURIER SERIES
identities,
e'
sectionally continuous in
where
may
Problem
cos
sin
e^*^
6,
written as
f(x)
Page
cos 9
sin 9
CnB^"'"'"'
(6)
can be
^
(7)
n= so
^l
where
Cn
w^ j
/(a;)e-"'^^'^da;
(8)
In writing the equality (7), we are supposing that the Dirichlet conditions are satisfied
If f{x) is discontinuous at x, the left side of (7)
f{x) is continuous at x.
1,
,j u
A by
V.
should
be replaced
1
f(x
-^^
+ 0)+f(x-0)
-.
^
BOUNDARY-VALUE PROBLEMS
Boundary-value problems seek to determine solutions of partial differential equations
satisfying certain prescribed conditions called boundary conditions. Some of these problems
can be solved by use of Fourier series (see Problem 24).
CHAP.
FOURIER SERIES
14]
301
ORTHOGONAL FUNCTIONS
Two vectors A and B are called orthogonal (perpendicular) if A'B =
A2B2 + A3B3 = 0, where A = Aii + Aaj+Agk and B = Bii + Bzj + Bsk.
or A1B1 +
Although not
geometrically or physically evident, these ideas can be generalized to include vectors with
more than three components. In particular we can think of a function, say A{x), as being a
vector with an infinity of components^ (i.e. an infinite dimensional vector), the value of each
component being specified by substituting a particular value of x in some interval (a, h).
It is natural in such case to define two functions, A{x) and B{x), as orthogonal in {a, b) if
r A{x)B{x)dx
if
A vector A is
A'A = A2 = 1.
normalized in
(9)
(a, b)
{A{x)ydx
(10)
is
clear that
we can
= 1,2,3,
.b
^mi^)
'l>ni^)
d^
m^W
{11)
X
In such case, each
also normalized.
is
The equations
member
{4>Jx)Ydx
m = 1,2,3,
...
{12)
member
We
call
{11)
and
{12)
set.
<kjx)'t>ni^)dx
{13)
'-'a
where
8m, called
Kronecker's symbol,
is
defined as
if
m^
and
1 if
m = n.
f{x)
^Cn4.Jx)
a^x^b
n=l
series, are of
(14)
[CHAP. 14
FOURIER SERIES
302
Solved Problems
FOURIER SERIES
Graph each
1.
/(a) fix)
0<a;<5
Period
-5 < a; <
-3
|_;
10
fix)
-- -t
3
-20
-15
-25
~*~
Per iod
--5
-10
10
IS
20
25
Fig. 14-2
Since the period is 10, that portion of the graph in -5 < a; < 5 (indicated heavy in Fig. 14-2
above) is extended periodically outside this range (indicated dashed). Note that f{x) is not defined
at a; = 0, 5, 5, 10, -10, 1 5, 15, etc. These values are the discontinuities of f(x).
sin X
Jib)
fix)
TT
O^x^TT
< < 27r
Period
2-0-
a;
/( X)
'-
-3:r
""
r^
-r
-2,r
/
/
iv
3^
2s-
Fig. 14-3
0Sa;<2
g <4
4ga;<6
fO
(c)
fix)
= \l
Period
a;
x and
is
continuous everywhere.
fix)
Period
-12
-|
-10
-I
8
-4
-6
-8
-2
1
12
10
14
Fig. 14-4
is
is
discontinuous at
4, 8, 10, 14,
s>sm T^L
X.
k-TvX
2.
Prove
kirx
sin
C
J
kvx
kv.
da;
COS
kvx
Kit
-j^ ax
COS ?
ax
kirx
^ sm j^
kir
Li
-y dx
\_,
if
COS
--,
kv
-j svakir
KIT
fcTT
A;
1, 2, 3,
-; COS
( ^tt)
kiT
Li
j sm
KV
I
1
(Kir)
i\
a;
2,
CHAP.
d.
FOURIER SERIES
14]
Prove
(a)
cos
303
sm p- sm -^ dx
m=n
.L
/i,\
(o)
where
(a)
-^r
Sin
sm
Y
TIttX
7
'"^ttX
cos
^^o
^^
j^ da;
From
trigonometry:
cos (A
+ B)}.
Then,
cos
cos B
m-^n, by Problem
it
l{cos (A
1, 2, 3,
- B) +
+ B)},
cos (A
sin
B - A{cos(A-B)^
sin
2,
=
Similarly
irly
m =.,
m"n,
if
nirx
X.^rmrx
^sm-^sm-^do;
m = n,
If
nvx
mjra;
Note that
have
if
sin
m = m ==
cos
B =
J_^,,n
The
c,c
A +
show that
n=
(a)
- B) +
|{sin (A
/I
cos
\,
jdx
2m!ra;
h J
+ B)}.
sin (A
L,
f-,
(1
respectively.
Then by Problem
L^
gj,,!^''^
for
;^
J_^
2, if
sm^^jdx
w ?^ m,
=
and
-J_^
(6)
sm J da;
respectively.
2/
a- cos -j--
1,2,3,
/(^)
cos-^da;
+ 2L
If the series
(m + nWx^
cos^-^jdx
replaced by
m = m,
If
(a)
da;
1 C
I
-;^
^sm-^cos-^rfa;
4.
nirx
X.sm j^ sm
We
we have
Xmrx
(6)
(m njirx
\ C
-J_^jcos^-^
b sin
-j-
.. .,
cos"^
dx,
{b)
A +
K^^ j^^
fix)
sin"^
dx,
(c)
A = ^.
Multiplying
fix)
,
by cos
/(a;)
rriTrx
cos
da;
4-
=
Thus
om
to
Z/,
using Problem
3,
(i)
we have
cos-jdx
OmL
(^acos^ + 6sin^)
-^ j
"S!
J
<,an
r
I
(^g)
^o ^'^^ cos
cos
^"a;
-j^ dx
+
1
r''
ifm#0
f(x)
cos^^^
da;
if
m=
1, 2, 3,
cos
1
= sm rura;
-j- dx y
mjra;
[CHAP. 14
FOURIER SERIES
304
(6)
Multiplying
mtrX
sin^^
by
(-Z)
-L
LJ
lllLegXai/lllg' from
XrUIIl
ana integrating
and
Wi=
AUUlCiU 3,
XJ, using
UOmg aProblem
O, we
to
l-U L,
4lv:;
have
Ld
I
sm-jdx
fix)
Thus
(c)
sin
=
I;
(1)
from
r'
Putting
m=
The above
sm j^
cos
f(x)sin^dx
Problem
to L, using
/(x)dx
i a
(S)
-J-
dx
bn
sin
= sin -= dx >
bmL
Integration of
^ ax
= 2AL
when the
1,2,3, ...
we
m=
gives
2,
or
if
"^
find
C''
f^^^*^^
oLJ
f^'^^ ^"^
J^
^^^ so
A=
integration limits
Note that in all parts above, interchange of summation and integration is valid because the
assumed to converge uniformly to f(x) in {-L,L). Even when this assumption is not
warranted, the coefficients a and 6 as obtained above are called Fourier coefficients corresponding
series
to f(x), and the corresponding series with these values of Om and 6 is called the Fourier
corresponding to f{x). An important problem in this case is to investigate conditions under which
series is
5.
(a)
coefficients
A)={3
(6)
(c)
How
of f{x)
shown
is
^"^'^-^^
"0<:<5
series.
= -5, = and
5^a;S5?
a;
a;
=b
Period
3
i
.....
1
-5
-10
-15
15
10
Fig. 14-5
(a)
Period
= 2L =
.
10 and
L=
5.
^""/(^)cos^dx
-(
sin-zo /
0,
5 \niT
lfn =
+ 2L
as
\fj{x)cos^dx
ifn^O
lo
IJ cos^ dx
-5
= ^
dx
3.
to 5,
so that c
= -5. Then
CHAP.
FOURIER SERIES
14]
305
Sin
3/5
5
(b)
In COS
=l
_
=
"
of continuity and
"
/(^
0)
a;
Expand
(a)
f{x)
a;^,
in a
wn-a;
-
sm
sin
.
5
.
SttX
(3
+ 0)/2 =
At
a;
-5,
and
5,
with period
f{x)
2ir is
<
a;
Period
= 2L =
""
Fourier series
shown
if
(a)
the period
is 2tt,
(6)
the period
o
2)r
4,r
6-
Choosing c
^.
= 0, we
a;^cosna;da;
i{(.^)(^) - (2.)(:i^^) +
x'
have
=
bn
L=
f{x)cos~-dx
LJ
=
If
and
2n-
dx
2(r)
dx
a;''
sin
f(x)
x^
-T,
n-0
nx dx
'=^'^^^-(2.)p-i||^) +(2)
Then
/(x)sm
-^J^
5 +
we
10
Fig. 14-6
Period
If
a;
-2ff
-4s-
which are
x=5
5 S g 5.
r^
points
= -5
a;
/(af)
-6i-
all
3/2
not specified.
The graph of
0<a;<5
3
3/2
is
s-a;
^^ p^j^^^^ ^^ discontinuity.
3/2
L
cos
wtt)
z^j^^z:l^
1
+ 6 /
3s-a;
+ S^'^T
+ ^(^''^T
(.
we can
-5 <
f(x)
3(1
-^
~
rejT
3/2
6.
riTr)
series is
L y
Lj
cos
tlir
nirx \
6 sin -^^F^
p +
3(1
.i
n
nirx\
wti-
(c)
= da;
? cos nx
n
4s-
Sin
COSi
(
\
J
nx
-4ir
This
<
By
a;
a;
= 2b-
Ill
+ +
prove that
6,
j2
"
4
+ ^
2^ ;? =
and so
2,^^
^^
^n^
_
-
a;
TT^
4
rj-
^ + 2
-3-
32"^'" ^6~"
22
4^2
4:7=
and
and
Then
ODD
At a =
27r.
8.
<
valid for
is
If the period is
(6)
7.
[CHAP. 14
FOURIER SERIES
306
1(0
to
+ 4v') =
2ir^
^'
~&-
Classify each of the following functions according as they are even, odd, or neither
even nor odd.
mm
From
Period = 6
_ltltl
{_l
it is
-/(),
fix)
2
Fig. 14-7
Prom
it is
is
-27r
Fig. 14-8
j'(c)
f{x)
^ x{10-x),
From
0<a;<10,
Period
10.
iw
y
/
\
/
\
/
\
VY
>^^ ^^s^
X.
/
/
\
\
//
\
\
\
\
^5
\\
/
/
\\
/
/
1
-10
Fig. 14-9
10
CHAP.
9,
FOURIER SERIES
14]
Show
Method
Fourier expansion.
its
1:
No
6.
If
3O7
we make
0,
~f_J{x).in^dx
the transformation
i/(.)sin:^d.
= -m
a;
on the right of
(-^yu
fi-u) sin
-JJ['/Wsin^du
j^
Then
f^i-u) sin^ du
(.)
and
in particular x.
|^
f(-x)
(a,cos^ - 6.sin^).
cos^ +
6,
Thus from
(1)
sin^^
Hence
f{x).
2 6sin^
^'^d so
obtain
cos^ + K sin^) =
(a.
we
f(x)
Assume
2:
(1),
(1)
-ij''/(a.)sin^ dx
where we have used the fact that for an even function f(-u) =
f{u)
dummy variable of mtegration u can be replaced by any other symbol,
using (2), we have
Method
us write
this, let
j^fjix)sin^dx+Lfix).m^dx
hS^
To show
1, 2, 3,
0,
f{x)
i.e.
+
=
(a.
cos^ -
^+ 2
a.cos^
sin^)
In a similar manner
Fourier expansion.
()
"^
--
Letting x
we can show
show that
Z Si
(a)
= u,
^fj(x)cosr^dx
an
/(^) cos
dx
^J
=
i /%(-.)
(6)
11.
j;
f^
Expand
f{x)
sin
a;,
f{x)
cos dx,
fl /(x) cos^
COS
(=^)d.
f(u).
Problem
< <
a;
tt,
+ i /'
0.
/('>')
cos^ dx
"
^f"fiu)cos^du
1 J^' /(;)
Then
'^^
(b) 6
eos
dx
9.
so that
L = v.
[CHAP. 14
FOURIER SERIES
308
f(x)
/
\
Y
2)r
-2ff
Fig.l4-H
By Problem
6n
10,
r"
H
=
{sin(a;
-2(1
and
f{x) COS
+ na;) +
%
sin
Ni
na;)}
sin
(aj
cos nv)
cos
nx dx
1 J
--!
+ l)a;
cos (w
cos (w
1
r-^
m+1
- l)a;
n-1
n^l.
if
-1)
a;
da;
n-1
+l
^^ dx
For
Ji
1,
di
sm X
cos
For
n-d.
(U,
sin
da;
a;
2sin'a;'
aa;
a;
= 2 ( cos x)
"^
1-
Then
2|(l +
IT
TT
cos.)^^^^^
n^
=2
4 /cos
2a;
-1
cos
cos
4a;
6a;
TT
12.
Expand
(o)
f{x)
Extend the
<
x,
a;
< 2,
in a half
range
sine series,
(a)
(b)
cosine series.
Fig. 14-11
Thus
On
6.
and
^(^\;^*=^-2-;
a;)
sin
"
Then
/(a;)
n=
_4
nV
dx
-4
nirx\
nirX
n-a;
sm-g-
K-
^-^ Bin
(1)
nTT
nirX
sms
.
da;
2
2!ra;
Zrx
- 2^">-2~ + 3^'"'2~
,
-)
cos ms-
shown
= 2.
in
CHAP.
(6)
FOURIER SERIES
14]
309
Extend the
This
is
/()
.^^s
\
Fig. 14-12
Thus
0,
= 0,
ao
Then
aa;
niT-
if
1)
n 7^
X dx
f(=^)
X cos
-J
^(^)(i^^"T)-(K;^-T)
-j-i (cos
If
f(^)<^o.-^dx
lX
+ n=i
247(cos,r-l)cos^
nV
jra;
- -^eos- + 3,cos^ +
It
in (a)
and
377-a;
f(x)
5,73;
g,cos5|^
= x,0<x<2,
+
is
\
.)
(6).
PARSEVAL'S IDENTITY
13.
fix)
assumed
is
^f
=i\
_i_
to
um^ dx
r'^
(which
^dx
nirx
f f_J{x) dx
^L
+ Mal +
bl)
nirx\
^'"TL~y'
to exist.
~L~
is
uniformly convergent)
{a fjix)
cos
d^
we
obtain
f(x) sin
dx\
(al+bl)
results
J_^/(x)cos^dx =
La^,
f(x)sin^dx =
Lb,,
f''f(x)dx
La,
(2)
The required
by L. Parseval's identity
is
valid under
14
1111
"
(a)
(b)
Determine from
(a)
Here
L=
2, oo
the
(a)
sum S
= ;^2
2,
all
(cos n,r
rt
1),
jT+24 +
^ 0,
6,
34"'"
'^v^'^
'"
'
0.
|o
M,
positive integers
Li xJ -I,
=l
S =
f'-on^^hich
l^+ S'
where
of the series
12(b).
identity becomes
Then Parseval's
15.
[CHAP. 14
FOURIER SERIES
310
corresponding to
coeificients
f(x),
and
f{x) is
assumed
For
Sm{x)
Let
M=
We
1,2, 3,
-j
+ bnam-j-J
facos-j^
sums
(1)
/^t
{fix)
f{x)SM{x)dx
Multiplying both sides of
2
f(x)
(i)
limit as
and
Sl{x)dx
we
,
2
If the equality holds,
from
M^
we have
Sm{x) dx
(2)
SM{x)dx
{f{x)y dx
(1)
SMix)Y dx
^l
13, gives
Substitution of
Taking the
2,^
have
Problem
(2.
2l\^ +
2^
-L
to L, using equations
{al
of
l||- + 2^K+6S)|
Problem
(4)
6?)
to L, using
(2)
-L
and dividing by
{3)
3,
we
find
(5)
n=l
J-'
^-L
can think of Sm(x) as representing an approximation to f(x), while the left hand side of (2),
indicates
divided by 2L, represents the mean square error of the approximation. Parseval's identity
^ o the mean square error approaches zero, while Bessel's inequality indicates the possibility
that as
that this mean square error does not approach zero.
We
The results are connected with the idea of completeness of an orthonormal set. If, for example,
could never
were
to leave out one or more terms in a Fourier series (say cos Attx/L, for example) we
we
analogy
get the mean square error to approach zero no matter how many terms we took. For an
Problem
60.
vectors,
see
dimensional
3
with
CHAP.
FOURIER SERIES
14]
DIFFERENTIATION
16. (a)
INTEGRATION
12(a).
(a)
and
(b)
Use
f{x)
to
a;
To determine C
< a; < 2. Then
in
L=2
since
...)
(2)
n=i
(2)
in this case,
we have
in (a),
2=
%'
16*3
4*
3^
12
Term by term
differentiation yields
(cos^ -
cos + cos -
CONVERGENCE
18.
ttX
1
2vx
16 f
1
Stx
^{cos-^-^co.^
+ ^cos^-
n
C -
Iffi-l + l-i.-..^
C =
where
Show
find
^2
X
17.
^ ^"H"
"
12(a),
(b)
w^ 1
From Problem
we
FOURIER SERIES
of
= x\ 0<x< 2, by
(a)
311
Prove that
of
(a)
not valid.
is
-..V
0,
FOURIER SERIES
i
cos
cos 2
sin(M + i)^
cos Mt
2 sin ^t
<b)
^' ^Jo
(a)
We
have
sin(M + i)^
2smit
cos nt sin it
l)
l{sin (n
""^
_
-
=1
to
in
2'
vrJ-.
sin (n
2'
- l)t}.
cos
Mt}
(sinft
(6)
,.
"^^
2sinii
M,
=
On
(M + i)t
sin
sin^t)
^{sin(M + i)t
(sin
(sinft
(M +
^)t
sinfi)
- sin(M-|)i^
sin^i}
fix) sin
nxdx =
lim
f{x)
cosnx dx
if
f(x)
is
sec-
tionally continuous.
This follows at once from Problem 15, since
lim a = lim 6, = 0.
if
n-+oo
n-*oo
The
Riemann's theorem.
the series
^^
=i
(al
b^)
is
convergent,
20.
[CHAP. 14
FOURIER SERIES
312
Prove that
We
lim
f{x) sin
M-
(M + l)x dx =
is
if f(x)
sectionally continuous.
have
+ i)x
dx
cosMx
da;
{/(a;)
J_
Then the required result follows at once by using the result of Problem 19, with f(x) replaced by
continuous if f{x) is.
f(x) sin ia; and /(a;) cos ^x respectively which are sectionally
are
a and 6 instead of ir and jr.
limits
integration
The result can also be proved when the
21.
Assuming that L =
2L = 27r, show that
Sm(x)
tt,
i.e.
S (an cos
nx
+ K
sin
nx
f^j'
r""
I
Then
SM(a;)
f{u) cos
18.
/W
cos n(M
(ft
Ji-
/(m)
dM
i r
f''
i
sin
/(m)
t,
tt,
+ x)
f{t
^^g
'^^
sin if
we have
+ (- j _
cos na;
Ma;
x)
sinnu sinna;j
/() sin
nw dwj
sin
dtt
du
cos
na;
+ i
+
(M +
.
6 sin a;)
/(w) cos
n(u
- x)
du
cosn(M-a;) ^dM
1)(m
- x)
; du
2sini(M-a;)
we have
1
sin(M +
r-^-'
A)t
Since the integrand has period 27r, we can replace the interval -jr
Thus we obtain the required result.
2jr, in particular -jr, tt.
a;, tt
a;
of length
22.
Prove that
From Problem
21,
1
sin(M
r"
+ l)J
5^^-^
Subtracting
(2)
from
(i)
nx
iX/^"^'^'"
/()-]|
ttJ-^'^'
ux =
Letting
L=
with
nu du\
/(m)
b sin nx)
coefficients
Also,
using Problem
na;
If".,
- 0)
2smlt
and f(x
+ 0)
jrj
sin(M +
l)t
respectively,
.i
sin ^t
CHAP.
FOURIER SERIES
14]
23. If f{x)
and
f'{x)
^^^
eontiJuouJ""'"""
^'^'
Thu<i
""^
/(<
+ a^) -
/(a:
+ 0)
'^
g sin ii
JimSW.)
-{
/<-
'
'''""""''"^ ^ ^'^^^
22,
^'^^
gtg
sectionally continuous in
because /{x)
1__ ^
2 sin If
sectionally continuous in
Similarly,
inO<tS^
+ -)~n- + 0)
^.
'^
2 sin ij
f(^
lim
,_+
" "'""""^
tis'ts"''
fii^:0)
sectionally continuous
}lZ- ^''^l\7jr'^
'-0+
Zsm^t
''''''*''"" ^'^'^
eacf;
^'
prove that
.r),
+ o) +
fi^
lim s.[x)
(-:r,
3I3
,,^^
is
nt + .)-f(.
,"?+
-^'^^ '^-'l
sectionally
+ 0)
d--ative of
/(.)
at
,r.
-tt
<
0.
we have
+ /(-")
or
= /(^+ "> +
Jim^(.)
/(^
- 0)
t)
_ 32c/
- ^aF
C/(0,f) = 0, C/(2,i) =
C/(x,0) =
at/
i>o,o<x<2
-Qt
i>0
0<a;<2
a;
mWrP
Fohht
^(^^) = 3(^r)
(')
(2)
and T
in place of
or
X(x) and
X^ = 3T^
dx
(^)
^^
dt
T(t).
can be written as
J_dr
3T dt
d^X
dx'
(^)
In Problem 47
we
see that
if c
0,
are independent
pciiueiii, variables
vdriaoies,
two ori":r;'rernr'etu:ti:nr^'^''^^
f +3x^r
^"'^^^^"^ -''"'''
0,
-^
g+
-""''^
X^X
- -'' ^^- ^-
(^)
^''^^-
:=
(,)
solution
is
Cie-^'",
U{x,t)
where
and
are constants.
X =
A, cosXa;
Bi sinXa;
(5)
e-''^''(AcosXa;
Bsinxa;)
^e)
[CHAP. 14
FOURIER SERIES
^.
We now
and B so that
we must have
seek to determine
the condition
t7(0, t)
0,
g-3x't(^)
(6) satisfies
JJ{x,t)
To
C/(2, t)
B=
m=
0,
1, 2,
Substitution in
(8)
i.e.
0,
2X
Wjt or X
(10)
first
OT.
If
we now attempt
impossiblus^gS
jr^r-e
also
sofutLs
U(x,
(i)
sine series.
The
0)
<
a;,
see,
^ J
X
it to be
we find
havmg the form
-J.
2,
solutions
Problem
12(a),
V,
U^)
t
= 0,
0<.<2
(^^)
function f{x)
equivalent to the problem of expanding the
<
on placing
B^sm^
we
a.
J^Be- sin^
= x,Q<x<2, we
X
is
U{x,
the condition
This, however,
boundary condition
m.,t)
From
is
= B6^'''*-sin2n|
U{x,t)
of
we
(^)
now shows
()
0,
sin 2X
where
(^)
makes the
satisfy
Be"'''* sinXa;
Be-"^'' sin 2\
Since
To
A=
or
from which we
=x
<
for
see that
x <^2 into a
B =
'="-
^
/t
actually a solution, we must show ^h^t
a /ormaJ so^^tion. To check that (U) is
^f
of
justification
consists
proof
The
conditions.
relitial differential equation and the boundary
accomplished
be
may
and
series
infinite
ttLTy ter^ Srentiation and use of limiting procedures for
Which
is
by methods of Chapter
11.
an interpretation in the theory of heat conThe boundary value problem considered here has
conduction in a thin rod or wire located
=
is the equation for heat
duction. The equation
f
=
fcg
ORTHOGONAL FUNCTIONS
25. (a)
Show
sin^, cos-^,
forms an orthogonal
^^'^rtT'
^^^IT'
'
'
CHAP.
(6)
FOURIER SERIES
14]
gj^g
set IS
(a)
By Problem
(b)
(a)'
^
so that the
3.
3,
dx
^^
j\lfdx =
Also,
~,
2L
C ( ~k\ dx
or
set is given
dx
by
^cos^
MISCELLANEOUS PROBLEMS
26.
We
f{x)
"'
T:
1
- r"
l/sin(-wV
jr
ao
S'""'^
>"(*)
cos
{cos(a-)a;
2 sin
2:r.
L = .
na; da;
a-n
2L =
2. so that
cos (a
^ 0, 1, 2, 3,
is
even, 6.
..j
and
nx dx
+ w)*} da;
+ V]
+w /
sin_(a
_^
2a sin
a,r
^(a'
cos n^
- )
a,r
Then
cos ax
2a sin
4-
sinar /l
air
v;
2
COS
2q:
,
~U~^'^1^'"' +
27.
Prove that
sin
a;
a;
X' V
V
^
a;
,7-
in the
= g^I^(J: +
a
7rC0ta:r
- i
a
is
of interest since
it
-^
2
cos
3a;
{2^)y\
{3^y^
cosa^
This result
2a:
^F^2^cos2a;
x^
-7V
Let
Wjt
-J^+
a^-1^
=
^"
2a
a=-2^
Then
2a_
^ a'-3=
_^^
a^-l^ ^ a-2^
^
.
2a
a2_32
(i)
[CHAP. 14
FOURIER SERIES
316
S \a\ \x\<\
converges uniformly for
test, the series on the right of {!)
the Weierstrass
Thus
s rule.
L'Hospital
usmg
by
seen
is
as
and the left hand side of {1) approaches zero as a-0,
obtam
a;
to
to
from
of
(1)
we can integrate both sides
By
^'""'^
ln(
In
so that
sin.x
vx
Replacing x by xlv,
we
lim
'"^liTil^-M^-l
In
_
A_^V1_^^...r1-!^
-^s)M^)-0-s)
=
=
i-^
(^-s)(!-s
Ci_e^
...
(.>
obtain
^'^
sin^
^(^-V^K^'i^)'"
TT
28.
Prove that
Let
a;
= 1/2
'
1-3
3-5
in equation {2) of
5-7
7-9
Problem
27.
Then,
(-^)o-i>)(-^)-
sides,
we
^ a-i)(i-i)(i-o--
is
^^^^-
^^^
FOURIER SERIES
3^^
Supplementary Problems
FOURIER SERIES
"
^^^f^^^-{-l
(c)
*^-
trol",:ttlt^^^^^^^
/(a;)
Period 4
ltr<l
0<a;<10,
4a;,
{-
(.)/(.)=
Periodic
(d)
f{x)
^^
(0
Ans.
(a)
31.
"
32.
(a)
a;
(c)
a;
Expand
^^
^^^'^
sin^^
(6)
^ 2 ^ sin^
20
(c)
Ans.
i^
= 0, 2, 4, ...
= 0, 10, 20,
.
/(.)
(d)
-x
<
20
(d)
<
a;
Period b
(l-cos^)
,1^:,
- + 5 / 6(cos^7r-l)
^a;
.^
3, 9, 15,
6 cosn^
^a;l
^ ^^^^.^^ ^^^.^^ ^^
(a)
Expand
(6)
How
/(x)
<
cos x,
a;
<
^^^.^
^_
rL^tTvrl^^^r'Li:^'"'
in
:r,
a;
and
m=
a;
cosx
Ans. Answer
Expand
Ans.
(a)
f{x)
is
the
same as
in
Problem
-J8_^ 4<x<8
^2^ f
sin
Prove that f or
sin
,r,
:.
0<x<.
35.
S <3
-3<a;<0
,^^^^^^
''
34.
J. JO
no discontinuities
(6)
2-^^
-4
in a series of
(b)
^-(^ +
(6)
xW-x)
f(li^+^ + 5ii5._^__Nj
37.
Show
that
if
is .;
and
(6)
any.
show that
~+~-^-l + L + l__
(6) cosines.
-^cos.
i^ +
to
(a) sines,
^ % /2cosW2
xW-x)
the period
32.
(a)
36.
if
and differences
3,r^v/2
A ^^^ ^
iorQ^x^n->
[CHAP. 14
FOURIER SERIES
g^g
FOURIER SERIES
of
tt,
By
(6)
a;
(o),
By
39.
Show that
(o)
Use
(6)
By
<
\
)
-,r
/sin
12(-p
3^
s in 2a;
a;
cos 3a
25~
sin
2^"
"*"
3a;
"j
3'
ir,
3454x
^sin* + 2(j43sin2a; - a^lsinS* +
(a) to
-tt
sinx
as
40.
as
(6),
- a;)(7r + x)
cos X
a;
<
for -a-
cos2x
Vcosa;
tt''
(c)
3a:
= -^ ~ ^\^V~
*'
sin
2~"
^-j
^3
sin 2g
/sina;
tt,
cos
Problem
a;
35(6),
/ cos 2a; _
\V^
"
cos 3x
cos3x
lr\^ +
\
J
tt,
\
V
5^
3-5
"^
S
cos5x
3^
cos4x
2-4
4 /cos X
tr
PARSEVAL'S IDENTITY
41.
By using Problem
2^;^ =
(a)
,n
42.
1
^^^
ot.
that*
Show *v,
43.
Show that
(a)
44.
Show that
^^.^i.
2^ (2n
1)*
+
+ ^^-^ .
_L
"t"
5577?
96
'
= 5'~^
^,^
(''^
90
[Hint:
ig
^'
=i (2%
945
Use Problem
11.]
960"
-D
39
+ 2^T^^
3.
"^
F^l^^
"^
"
16
BOUNDARY-VALUE PROBLEMS
45.
(a)
^=
Solve
where
<
a;
< 4,
>
1/(0, i)
=^
0.
V{x, 0)
8 .in
2 sin
Ux.
and
solution.
Solve
0.
46.
0, 1/(4, t)
1/(0, t)
0,
t/(6, t)
0,
U(x,0)
= jj 5J*<6
Ans.
47.
t/(x,t)
.2.
Tl - cos(W3)1
cos (W3) ^_2^236
Jl
J^2|_
J'
Show that
if
'"'^*
^^;^
is
a constant c where c
S 0,
then there
is
no solution
CHAP.
14]-
FOURIER SERIES
'"'
^^^--"
. = and . = . o the . axis, its ends
transverse vibration the displacement
Y{x, t) from the x
ttr^'l^::ti:2'Xl:n'^^^^^
points. When set into small
r^^
=^S
StllS'^"^"*^^^"'"^^^^^^ ff
'^
Where
the
fi1i^o:rm?='o\7:T^"'
"' ^('^'
^(^. 0)
0'
*)
gjg
sm X +
0.1
....
a^
r/,.
r = tension,
e,a.io) with a^
=4
for
= mass
which
<x<
per
satisfies
>
^, t
o.
(6)
49.
(a)
Yix, 0)
Ans.
50.
(a)
- .),
0.05.(2
Y(x.t)
Y.ix, 0)
g^ =
<
where
0.
^ 5 ^^^
51.
Z7(a;, <)
f = g,
2^+5
< 2. * > 0.
1/(0,
t)
'='"' ^^^'^
4_cosm^
(6)
If
~^<x<v
and a
7^ 0,
If
< <
77
a;
jT,
t7(.,
1,
"
'' *
*)
1, 2,
r(2.
U(x, 0)
*^ '^^^^^-^-^
^-^-"
50.
Prove that
sinh
Prove that
cos
Show
a;
a;
a;
(^1
(l
that
a;
=:
(a)
a2
irV2
_J_
2 sin 2x
2'
- a^
a"
3 sin
'^
-ii^
a cos
a;
+V
Vl
a"
32
3a;
_ 2
^'
,
^x^
a;).]
11
13
3 sin
2
2a;
+ 2^
(3.rA^~(5^
15
2-2-6'6-10'10'14-14
2a;
+ 22 +
^LV,
cos
g.
^Y 1
_ i^
2 sin
;^2
J_
2a
-^A^
,
(sm
2x)/(2 sin
(6)
a;
sing
V
cos
sin
V -a'
aTT
cosh ax
2 sinh OTT
*)
2.
,,
prove that
.,
;r
(6)
r. ,
[Hint:
ajT
5sinW
2 Sinh
58.
3,
^^"^^
prove that
(a)
57.
0,
2 sin
56.
gi^^^sin(?^L^sin2(l2i^
5Sina
55.
*)
Interpret physically.
MISCELLANEOUS PROBLEMS
54.
sin ma;
An.. F(.,.)
53.
=:
Eary1onS;nsL''yt^;]''^'' "'
Ans.
...
i/ 4.4.8.8-12-12-16-16... \
V3'5'7-9-ll-13'15-17..y
3a;
+ 32
_
-<!
[CHAP. 14
FOURIER SERIES
59.
Prove that
(a)
if
a' 0,1,2,
.,
then
2a
Prove that
(6)
<a<
if
_|^
2a
_ +
1,
dx
2a
(c)
60.
From
(a), (6)
and Prohlem
17,
Chapter
13,
_2a
2a
Show that
{T'iY
(r- jf
(r)S
('
('"
J)'
61.
w=
a minimum when
If
_,_
{0()},
1,2,3,...
^'^'^'
c,
Hx)4>Ax)dx
series.
Discuss the relevance of this result to Fourier
"^
identity.
{f(x)
is
nI
Cn^.Wrd*
is
chapter 15
Fourier Integrals
FOURIER INTEGRAL
The
2.
\f{x)\dx
converges,
i.e.
finite interval
fix):
j^
(-L L)
oc).
states that
{A{a) COH ax
^()
Bia)
If
fix) COS ax
where
The
dx
^^^
fix) sin
ax dx
wo must
the
aho.
The
Sht
replace /,)
by 0)ifc<>)
conaitions are
sucient\t not
If
i.
.
,
" "'' ' "-""""""'y.
..
^^ ^ """"
lei:
^^^^
''"'^
'"'
S^^l-SrrS-^-^---- ~^
EQUIVALENT FORMS
of
FOURIER'S INTEGRAL
THEOREM
"
'^
X=o
J= _
^^""^ '^"^
"(^
~ ") ^ <^
(3)
and wetaVe""'*^
fix)
^^''^
=
=
^S^
cos ax da
'^" "^
^'^
if /(.) is
aUdu
f^
fiu) cos
321
if /(^) is
even
^f /(=^) is
odd
(5)
t'^^^P- ^^
FOURIER INTEGRALS
322
FOURIER TRANSFORMS
From
follows that
{A) it
if
then
r./
n^-^
:^jy^"^
Fmjripr transform of
+V,*.
^ Tjf,;^rT^tl:.^i^li^":^'^^''^'^^r
is 1/2..
product
If f{x) is
The above
is called
fix)
and
trlJfor^
is
of
sometimes written
^M
and
is
w-itten
^'^
'""''''
(5)
yields
^||J^(u)c0Sl.d
^^^
/(a;)
and we
call i^c()
If f{x) is
and we
and
and
-Fsia)
/(a,)
/(x)
G.W
and
(6)
if
F.()
^^^^
Fs()sinxda
-\(|J^
other.
i^oner sine trans/orms of each
for
FOURIER INTEGRALS
of
are Fourier sine transforms
/(.)
and
==
.(x),
HD
and
(i^)
/(.)
and
.(.);
then
become respectively
{fix)Vdx
i^')
f{F4a)Vda =
j\fi^)Vdx
(^^)
relations are
Lown
as
^--fV^/,^^^^^^^^
F{a) and
Fourier transforms. Thus if
tions hold for general
prove that
0(x) respectively, we can
and
(^^)
f{x)g{x)dx
f{Fs{a)rda
of fix)
,.
,
,^
then
.(.) respectively,
J^
tralsfonns of
and Oew"are Fourier cosine
/(I)
In the special case where
The above
yields
>f|J^()sini.d^
fF.{a)Gs{a)da
Similarly
Fc{a) COS ax da
other.
Fourier cosine transforms of each
fix)
PARSEVAL'S IDENTITIES
F.W
-\||J^
call F.(a)
If
(x()
CHAP.
15]
FOURIER INTEGRALS
The
signifies the
F{a)G{a)da
%/
00
323
f{x)]g{x)
)
dx
^g^
by
-i.
CONVOLUTION THEOREM
If Fi.)
/_F{)G(a)e--d
If
we
"
and gix)
respectively, then
j{u)gix-u)du
(^5)
^*^
/(.)
Z^**
:^^J_.^(^)^(^-^)^^
(17)
J if* 9}
transforms.
^ran^/orms.
This
J{f}J{g}
^^g^
is
Solved Problems
FOURIER INTEGRAL
The
1.
(a)
(b)
Graph
(a)
/()
and
its
of f(x)
/:
For
tt
The .raphs
0,
we
of /(.)
21
o
-3
-2
Fig. 15-1
fix)
= I^
'
\0
giaa
FOURIER TRANSFORMS
and
obtain
and
F^
<a
|;r|>a
= 3.
is
F{a)
|a;(
g-iao\
,or a
V^
=B
;2 sin
ei
V^
la
aa
a.
Use the
(a)
2.
[CHAP. 15
FOURIER INTEGRALS
324
F(^
= -i- f"
if
then
dM
/(M) e'
The integrand
and (2), we have
^"^
"
!xl<a
1/2
lo
1 f"
r" sinaa_cosa ^^
\x\
|a;l
>a
sin
aa cos
the result of
1 in
^^
^-=0
and a
gg sin ax ^^
TT
l^l
<
^/g
||
a;
>
io
(a),
(2)
is zero.
Then from
we have
sm a
r^ilL^da
sin
f"
FWe-'^^da
The
If X
-7=
rl
r J2sinaa^_,,^^
V2^ ^-
(b)
/(a;)
1,
_1_
(i)
sin aa~cos aX ,
f"
J_^
evaluate
1 to
From
(a)
Problem
result of
or
TT
J
d
__
'a
3.
If
/(a;) is
(a)
= VIX"/(^)'=os^'^^'
F{a)
We
have
(a)
(1) is
^^^-
V2;^-^-
Xu
f(u) cos
au du
-Vl
(a),
/(m) cos
integral on the
au du
From
odd.
is
is
-^
F(a\
vi^J-
If /(z.) is even,
right of
(6)
^W^^^'^^^'^'
'V^X
"
^"^
^^""^
(^)
the sine.
4.
Let
sll
f(x) COS
/(^)
ax dx
f(,x)
cos
and choose
F(a)
^|lj'
F{a) cos ax da
of'
a)
5r
Jo
(1
axdx
cos
aa;
da
FW
=
_
|1 "
^7 /
2(1
^^^
\f
cos x)
-^
'^^
''ill'
^^ "" "^
^^-^yP-^-^'
^^^^-
FOURIER INTEGRALS
^^J
Use Problem 4
5.
to
show
f"^
that
2'
,2
As
obtained in Problem
325
4,
COS X
cos
-i
V
(
aa; da;
'^
-I
n<<i
O^aSl
"
a>l
find
J
But
this integral
C" 2
can be written as
COS
6.
= 2m,
Show
dx
sin' (x/2)
Jo
X
a;
a;*
^^'"^ becomes
^"^
x^
~^ du
f^sin^u
J^
on letting
r^^d.
+
that
'^0
Let fix)
a^
''^*'^
"
-J^
But by Problem
X^Q.
2^-^
22,
Chapter
12,
cos
'^"^ a;
^ J
aa;
we have
da
da
awdw
e"" cos
/(m) cos
f^-cosattdM =
/o
\u du
e-
-^_
+1
da
Then
= e~'
PARSEVAL'S IDENTITY
7.
{/(a;)}2da;
where
/(.)
This
is
Wj"
J J
,,
i.()
{F{a)r da
V^?m^.
equivalent to
/ W'dx
r^^da
=
ri?}I^aa
=
2
-'-o..
f"?!!!!^ da
Jo
r sin'
i.e.,
Jo
By
'
letting
used to find
J^
= and
r^du
using Problem
directly.
da
a'
5. it is
a^
is correct.
PROOF
8.
[CHAP. 15
FOURIER INTEGRALS
of the
limitmg
Fourier's integral theorem by use of a
Present a heuristic demonstration of
of Fourier series.
form
Let
where
a.
= 1
Then by
If
oo
fix)
/W cos^
-^
and
du
11705
/^
(^ancos-^ +
Ksm ^^^
^
.
(1)
jjM sin^
= ^
6.
du.
we assume
/"
that
\f(u)\
du converges, the
first
(.)
approaches zero as
approach
L-^o, while the remaining part appears to
2
^ n=l
lim
L_
is
Aa
Li
r_
fiu) cos
y^{u-x)du
^
(3)
heuristic.
not rigorous and makes the demonstration
t/L,
{$)
can be written
f(x)
lim
Aa-*o
n=
^ ^( ^)
But the
limit
U)
is
is
- x)
CF(a)da
1
-IT
Chap.
To be
COB
rigorous,
we
a(u-x)dx
-^-oo
is
12.
r sin
Let .
Cm
fda
Jo
method
and examine the convergence. This
10.
du
i"*^^"^^'
(6)
- x)
If^daj
29,
(5)
du
equal to
Hx)
which
i r
f ()
-!/.
Then
at)
J_^-ir-'^"
if
lim
F{x)
is
]=o J
^Hlldt,
^
= -.
^
sectionally continuous in
F{x) sin
aa; da;
{a, b),
then
(see Problem
with a similar result for the cosine"
31).
Use
^"^^-
FOURIER INTEGRALS
^^J
where
/(x)
Using Problem
(a)
9(a), it is
{f{x
+ v)~f{x +
(0,1,)
proof of this
is
since
_^lim
We
F(^)
and
exists
:=
|/(.
+ 0),
F(v)
^'^'
/(^^
+ ^) -
f(x
"^
+ 0)
sectionally
^.
IS
we make
{a) if
i Sj(- + -)~^<^v
()
0) respectively.
0)}^^^^l^dv
and {-L,
in (0, L)
lim
(6)
32^
lim
ib)
use of Problem
|/(x)|dx
fix
9(6).
converges,
prove that
+ v)'J^dv=
Ifix
- 0).
have
/(.+.) 115^
r/(.+.)^d.
r/(.+.)^
dv
(1)
dv
(2)
Subtracting,
sin
at)
(3)
in (S)
by /,//, and
J^"/(.
+ .)^d, _
respectively,
we have
j?/
fix
= L + ll+I.
+ n\
0)
I
sin
ai)
d-y
so that
Idi;
Also
<
s
l/J
l/r^_um:
\f(x+o)\
f
f'
sin ai;
^^
Jince j;
!/(.),
d.
and
we can
f '-^
dv
both converge,
e/3.
Also,
we can
Ti,^.,
*^
choose
//\
^'^
12.
We
Since
^fjiu) cosa(x-M)
du\
y()
COS
(.-.) du
"'
so large that
i.
"^""^
in
"'
\fiu)\du
~ u)
du da
./3,
''""' """^
(a)
'
|7.|
f(^
which converges,
+ 0) +
it
on Page 321.
fix
- 0)
all .
Thus we can
[CHAP. 15
FOURIER INTEGRALS
328
- (
da
where we have
Letting
Lett
/(m)
^(^)
du
cos
SinL(u-a;)
o:{x
- u) du
^^
a;
u = x+v
let
i->
-u) du
we
==,
see
f(x
^
by Problem 11 that the given integral converges to
+ 0) +
/(a;
- 0)
as
required.
MISCELLANEOUS PROBLEMS
13.
^=S
Solve
U{x,
t) is
0, t
U{0,t)
U{x,0)
0,
Jq
y.^^,
> 0.
We
B{\)
e-'^
'
sin \x
d\
J'*
where B{\)
is
undetermined.
By
sm
AK aA
B(X)sinX.dX
ii(K}
we have
[iQ
'^lt\
s1
a;
(^)
B(\)
,.
f{x)
U^^^t)
is
2
-
sm \x dx
r'
_
-
sm ,\x dx
,
-cosX)
2(1
-^
(3)
given by
14.
Show
that e'"^'^
Since
e'''^'^ is
Letting
is its
32,
Chapter
4.r e-^cos(aV2)dw
which proves the required
15.
by
^/2U
e'"'''"'
becomes
^.^-^ =
cosxa dx.
e-^"
result.
y{x)
where
g(x)
and
g{x)
y{u)r{x~u)idu
I
Suppose that the Fourier transforms of y{x), g(x) and r(x) exist, and denote them by Y(a), G(a)
and R{a) respectively. Then taking the Fourier transform of both sides o the given integral equation,
we have by the convolution theorem
Y(a}
G(a)
^f2^Y{a)R(a)
or
Y{a)
=
1
- V^-B()
CHAP.
FOURIER INTEGRALS
15]
Then
y{x)
assuming
y-
-1-
329
GM
Supplementary Problems
The
FOURIER INTEGRAL
16.
(a)
and
FOURIER TRANSFORMS
fix)
^^^^
1^1
'
|a;|
(6)
.
17.
18.
sin ae
(6)
Evaluate
/(x)
( "^"^^^r
^
I
f(x)
ir
fin<^
19.
V|(^-^^)
(a)
the
(5)
Show
Ans.
\/27^
(a)
| 1
" a^'
I'"!
<
~
fj
(a)
d.
its
(6)
e'',
0.
= I e--, m >
why
in (a).
the result in
(6)
+ a')]
J^
Y{x) sin
xtdx
J 2
gi<
a 2
(2
2 cos
a;
4 cos 2x)/vx
PARSEVAL-S IDENTITY
21.
Evaluate
(a)
J^
[Hint:
^-^,
sine
f{x).
transform.
[ff/(l
^f ^.
W^Explain from
20.
(6)
(a)
that
result.
) cosfd..
Ans.
""
- 0+
(a)
If
-'
>e
(6)j^"_^!^
by use of Parseval's
>
0.]
identity.
Ans.
(a)
W4,
(ft)
^/4
[CHAP. 15
FOURIER INTEGRALS
g3Q
22.
Use Problem 18
23.
Show that
show that
to
r"
(g cos X
^^
(a)
sin
Ze
ax
(^
^^
_
-
xf j
''3'
1'
0,
17(^.
t)
"*
^^
J^
77
15-
MISCELLANEOUS PROBLEMS
24.
25.
^.
U(0,
Solve
(6)
f = ^.
Solve
An.
26.
^=
(a)
(a)
Show
t)
lj;"(^
t7(..*)
17.(0,
-',
= {^
^(..0)
0,
^f
28.
Establish equation
29.
If
(i),
(18),
Page
Page
321,
from equation
Now make
^^f me^-du
the transformation
u+v
f(x)
(S),
From
(a)
signifies the
> 0.
.> 0,
G(a)
'^^
g(x)
321.
=^
^1
| ^
\x\<l
jail
>
'
ei''^''+'
F(a)
(see
e^"^
dv.
then
f(u)g(v)dudv
x.]
GW
- (14),
Problem
^^ fjM
da
f(.x)
complex conjugate.
Page
and
31.
0, t
323.
F(a)G(a)
(6)
>
bounded where
" 7=J
^=^
(a) satisfies
(a) If
bounded where x
is
^(^.
'^lt\'
e-"^'^^
27.
30.
>
+ ^^^)e-='cosX.dX
[Hint:
(6)
V{x, 0)
0,
Page
10).
322.
and
g{.x)
/(*)
f'*
13.
>
0.
chapter 16
Elliptic Integrals
The
tfLt u
^rst
first kind.
Hnd
by
Dy
F{k,4.)
^^nSf
'"^
''''
The
f ~Ji=^
FIRST KIND
of the
'"''^'
'^ '^^
-r^^Jt
will be assumed that
of the
form for
integral
is
(n
i '''
-^-^-' written
A;
SECOND KIND
Vl-A;2sin2^ de
also called Legendre's
0<k<i
it
defined as
is
is
defined by
0<k<l
(^)
The
THIRD KIND
do
C.
n{k,n,^)
of the
(l
+ nsin2^)v/l-A:^sin"^
<l>
= n/2
JACOBI'S
the integral
FORMS
for the
is
we
defined by
0<k<\
" <
<1
^?^
{3)
A:
kind.
Here n
is
a con
(1).
ELLIPTIC INTEGRALS
transformation
f^cvof above,
*K^
tegrals
is
dv
F,ik,x)
Ei{k,x)
C\[II^dv
331
[CHAP. 16
ELLIPTIC INTEGRALS
332
dv
ni(fc,n,a;)
/0
called Jacobi's
+ nv^W{-i--v')0kV)
(1
elliptic integrals
INTEGRALS REDUCIBLE
if
= l.
ELLIPTIC TYPE
to
and
x
If R{x,y) is a rational algebraic function of
nomials in X and
y,
then
y,
i.e.
-^
,.,,
R{x,y)dx
y^i
of elliptic
given constants, (7) can be evaluated in terms
functions.
elementary
of
terms
in
cases
kinds or for special
polynomial of degree greater than four,
If y=. ,/P(^) where P{x) is a
(7)
may
be
functions.
integrated with the aid of hyper-elliptic
,
^ =
. .v,
.
is
elliptic integral of the first kind
the
for
form
Jacobi
the
in
x
limit
The upper
it follows
am,
=
Since
=
<^
.
sin^
form by
related to the upper limits in the Legendre
= sin (am u). Thus we are led to define the elliptic functions
that
sn %
cnu
(^)
Vl-'f'sn^M
dnM
(10)
i!L!i
tn
iii)
sin(amtt)
^/l-x^
=
=
cos
^1 -
kV
(am
w)
as
to those of trigonometric functions
which have many important properties analogous
example, if
dependence
this
emphasize
To
= sn-l Note that u depends on k.
la;,
modfc.
write u^srv-^{x,k) or m = sn
elliptic functions; for
It is also possible to define inverse
= sn
then
sometimes
we
a;
LANDEN'S TRANSFORMATION
By
tan
4,
'^"^^^
=
k
X
where
ki
|^
(see
or
we can show
i
^l-k'sm^<j>
Problem
sin
fc
<^
sin
(2.^,
<#.)
{1^)
cos 2^j
61).
that
+ kJo
^l-klsm^.i>,
F{k,i.)
j^nK.'t>;)
(^^)
^"^^-
ELLIPTIC INTEGRALS
^^J
hm
K=
From
this
X ^TT^n^
1.
^
.
where
333
< 1 and we
follows that
it
^IZtlfT" ""f'
"
^-
^^ tan
(^j
+ -j
(,5)
Solved Problems
ELLIPTIC INTEGRALS
1.
Prove that
By
0<k<l,
if
-;)-'
';"l:;t:tL';'''""
'
""
""
(rl/2)(-3/2)
{-1/2X-X)
(-l/2)(-3/2)(-5/2)
,,
" "-"""
"" "'"
<'
"
,3
'"'''
b. ten.
1^1 +
using-
Problem
15,
Chapter
13.
7^1^
^'^
elliptic integral.
Let X
-Tr/2-y and
f"
^0
Let cos
1/
cos^ u.
^y
2 c os
sin
udu
= ~2
f-dy__
^
V^^
= y/Tin
p^^
-^o
=
Substituting k
sin
2 cos
^^"COS***
Hence
cos
dy
Vcosy
m dw and
sin
uju__ ^
Vl-COS^M Vl +
du
VI +
\/2F(x/I,W2)
r^n
^0
COS'
or
COS^M
g^
V2-sin'M
2 cos
Vl +
u du
COS='M
^,
X 7f^
V2K{y/^).
[CHAP. 16
ELLIPTIC INTEGRALS
gg^
Another method:
V2
Let
"
Vii^
Jo
sin yl2
V^^
Jo
sin 0.
V2cos0g^
J"
Vcos^2//2-sin^^/2
-^o
dj/
cos
4-
d0.
V^ f^
^fl-2^in^vl%
"!
I.
= '^''^'^^
^l
77/2
Evaluate
3.
Let cos X
/mz
v^cos^
f"
cos''
as in Problem
^'^'^^
Jo
^0
r^n
1
Jo
a;
2V2
2V2E(^/i)
7r/2
sin*
4 sin^ X
V4
Let
sin*
2 cos u du
Then
r'
4(1
C""'
i \
du
V2 -
2=
f
I
f" Vl + COS* u
""
sm* M
dM
'^'*
=====:
cos* .)
dx
V5-4cos*.d..
2 sin m
cos
in
sin
d0,
V5'Vl-tsin*,-d,
terms of incomplete
d0
cos
g cos u
_
-
dw
r*
J^
where
d0
cos
2y/\
^^^
sin*
g^
</,
COS*0d0_
C Vi-lsin*0
2 Jo
r*ziiii^:i^^^d0
-^/l
C"^
2 Jo
V5B(V475)
elliptic integrals
COS0
^
cos0.^-^===d0
Jo
_3
=
<p.
2/,
IVT^TT^i^du
where
cos*" dM
- ^f^K(^)
p Vr+
V'l
Jo
da;.
f VT^TiiSHT du
Express
C"
du
Vi + cos* u
-'o
j;"V 5-4sin*^ d^
5.
integral equals
Letting
udu
V2
"^
cos'u)-l ^^
~^
2
COS* u
{l
f" cos^ u
^^'2
VI+n +
CTrli
The
2cosj^\
'=^"(vfT^^)
r^'^
/I +
Then
2.
f^"
elliptic integrals
terms of
in
rfx
sin*
tt0
Vl - i
+
sin*
-fF(i,0) + 2E(i,0)
r VT-isin*"!^ d0
-^o
S x ^ W6.
CHAP.
16]
ELLIPTIC INTEGRALS
^^==
2
COS.
(cos^./2
is
= 3-2
sin^./2)
cos^./2
a;/2
;r/2-M
it
elhpt:c type.
6,
f"'
'^
from ./2
= ./2-.
t?^
\/2
v/2
-A ^f __*L__
"
becomes
seen that
it is
_2_] ^^../2
f"
f VTT^E^
V2-sin'xda;
ellipse
a.
VCd^FTw
a sin
d.
The
10.
(a^-6^)/^
oVa^
j;*-^_|__
Let fcsin0
^<*
tanw.
_
_
./2.
2/
2PvrT^^d.
Jo
6 cos ^, a
2V2;(^/|)
> 6 > 0.
6'sin^,
rf,
4aJ^'^"V l-e-sin-^ d^
terms of
in
Then k
cos ^
f" secM
d,/,
.,
elliptic integrals.
sec'
u du and
^-
sec^
f"
"Vfc-tan^,.
^"
-'o
V/bVos=M-sin^M
du
I
Now
^^^
proceed as in Problem
Upon
upper limit
retracing steps
it is
5.
Let
VF+l
sin.
A=
sin..
Then
VTlcosudu =
by
a;
^^'
"*"
nn-'(
Vvi +
^ ^'"
fe'sin^^/
"'
Then
is
Express
^,
4"V^eosV
Va^-(a^--6^y^hiF^d^
e^
and
Vl-^sin^tt J
2v^j/'Vl-isin^a;rfa;
where
y=^amx, OSx^^.
f'VT+md^d.
when
/W2
9.
'^
sin^M
Vl-fsin^t*
....
"'"'^
^^^fll^
Vl - |
-i/s J,
-v/3-',r/2
a;
VaVo
_ _2_
cos
V3^. Vl-|sin^M
8.
\/i-|cos^a;/2
V^
From Problem
since
^J
V3-2cos=~^
Letting
335
is
, cos. d.
related to the
11.
[CHAP. 16
ELLIPTIC INTEGRALS
ggg
elliptic integrals.
(a)
Letting x
2 sin
9,
*y
V9 -
^0
'^'^
(b)
Letting x
Vl - I
^ -^o
4 sin^ e
tan
sin= 8
e,
V +
Jo
tan'' e
tan'9
a/1
l
Vr+Tto^
Vl + 2 tan^ e
Vcos^ 9
Vcos^9
J_ r"*
Letting
ff/2
0,
1_ p'*
-'^M
V2
r'
<^a;
V^^^ =
tan
9,
V2
cos* 9
In general
2)(it*
e^^
transformed into an
if some zeros are complex
12.
Evaluate
sec 9
p/
{F(Vl72, ^/3)
V2
P,{x)
is
F(Vl72,
d9
r'"'"
"^"^
VT^I^ta^
jr/4)}
zeros,
a third degree polynomial with real
by the method
Problem 12).
integral
can be
indicated.
sec 9 tan 9 d9
V(sec*9-l)(sec*9
Show how
,
Make
1
+
to evaluate
f
J
de
C^"
-^o
3)
13.
jfl
Let M
Jo
~
=
(see
+ 1)
^J^ V2^^^i^
f -^^- where
integral becomes
Then the
becomes
this integral
sin'
F( Vl72, W4)}
^^
V(m'+
-3
.n
x^3 + uK
M or
^ Ji
{f (Vl72, W2)
V2
J^ V(x-l)(x-2)(x-3)'
Letting m
sin*
Let
2 COS* e
becomes
d,p
Vl - i
cos^ 8
de
Vl ~
-^|>
\/2
V2 -
2sin^9
2 sin^ 9
d9
p'*
2^ Vl i
= 0, 1,
>
-^o
respectively.
^T-
^ sin* 9
\f{^I2,
^12)
|k(V3/2)
sin* 9
^
^^
-^=========
x
^^
f"'
^JTTJ^^e
correspond to
in
terms of
elliptic integrals.
1,2,3
CHAP.
ELLIPTIC INTEGRALS
16]
This leads to
Using
Then
1=^,
= u\ we
if t
In general
transformed into an
V{m^
where
Jdx
We
P4(a;)
Vf(t
,^
terms of
in
Let
V{y'
(2a
2)y
pu, where
is
/?
= 3.
where a
is
a positive constant.
10}{y^
(2a
l)y
coefficient of u' in
u = {|, du =
j^^.
we
let
impossible.
transformation
1}
- 2a + 10 =
a'
+a+7
is
i.e.
fl*
'^
=9
du
l)(u'
+ U + 1)
dt
yj(t'
t
tan
l)(t=
+ 3)
as in Problem 11(6).
as an elliptic integral.
+ 10)(a;2_2a; + 7)
+a
l=y
dy
is
+ 9)(;S%^ + 3/3U + 9)
X
which
a*
Then
\/(a;'-2a;
is
Then
J V{u' +
Jdx
If
i.e.
a^
+ Sy + 9)
each quadratic
which
''^
I
is
method of Problem 13 is
Then the integral
dy
9){y'
p(
^~
which
can be
a constant.
Then
Let
applicable.
elliptic integrals.
+ a,
+ a'-2a +
V(/3'm'
yS
^^^
dy
or
is
is
/ Viv' +
Let y
=^
- l)(t + 3)
Then
1.
or a
so that
proceed as follows.
= d,c = d
1)(m^ + 3)
elliptic integral
;,
becomes
3d,
'^^
obtain
J ^/==
from which a
f|, 3=f
337
reducible to standard
form by
V(2/'
letting
+ 9){y' + 6)
\/6 tan
e.
u =
Letting
Jo
(3
du
Evaluate
16.
sece,
cos^)Vl
3cos'^
"
Show how
is
to evaluate
cos^)\Al
(3
Jo
VTTsT^?^
+ 3co?
^
+ cos^)Vr+^^os^
(3
^-^o
^ -^o
(1 - 8 sin^ e
./l^Tf^^e
a/1
1F(V3/2,W2) - in(V3/2,-l/4,W2)
=
where the second integral
^'>
Jo
17.
[CHAP. 16
ELLIPTIC INTEGRALS
ooo
= M^
let t
iu'
C
J
^
_^=======
-
Vl - f
sin^
terms of
in
r
j
elliptic integrals.
{t
1) dt
---^^^^====
to obtain
+ i)du
r_J(3.^1+iHj
r
du
J V(M^-l)(^ +
Then proceed as
sin^ e)
third kind.
the complete elliptic integral of the
Now
r
in
Problems 12 and
Letting
Another method:
x-1 -
du
4 r
_L
+
r
3
3)
(3u^
_*L
+ l)V(^l)(' + 3)
16.
becomes
dy
ELLIPTIC FUNCTIONS
18.
Prove
By
(a)
^{.nu) ^ cnudnu,
definition,
if
m =
^{cnu) = -mudnu.
de
f*
)
-'o
(b)
,,
-i/l
fe
fe'
Vl -
snw = sin0 =
.,
SI
sin'
e
sin
(am
cnM =
m),
cos
(am
cos
m)
Thus
since
(a)
i:(en)
^"' '^'
^=
dtt
^
-
= ^(eos.) = - sin
du
and so
^=
a/1
g
-
^^
k^ sin^
<P
= cnudn,
snuVT^H?^^,
sin^ = dnw.
= -sn.dn.
^"^^'
^^^
ELLIPTIC INTEGRALS
-^ (dn m) = - A;^ sn m en m,
du
Prove
19.
Prove
20.
(a)
339
Let
= - ^n .,
s (-.)
(a)
Then
Now
snM =
sin0,
sni;
e = -^ in
Thn sn
on (i,\
Thus
= - sn M.
(-U)
letting
M en M
en (-.)
(6)
= r
-A;^ sn
en .,
f"
sin (-0)
= dn u,
dn (-.)
(c)
= - sin
= - sn m.
it
becomes
Since
(0)
Since
cnu =
^T^-i^,
en (-M)
21.
Show
that
if
if
en (-.)
The
-r
^n M^c) dn (M
xv;
f"
N/r-A^sin^5
on the right
+ ^,
+ 2/i:) = dn M,
r'*'+'^
(^tf
J Vl
Jo
.
k^ szn^7
k'
sin^7
Vl -
X
-'^
Vl -
sn(u
+ 2K) =
cn(K
2X)
J*;
^'^
dn(K + 2X)
id)
tn (m
+ 2K)
k' sin^"7
or-
"'"
+ 2i^
in
sn(t.
cn(M
is
complete.
+ 2ii:) =
2A^.
d^p
am(M + 2Jf) =
Vl -
+ .) = _,i^ = _,^
= eos(0 + .) = _ eos ^ = -en^.
= Vr-fe^snM.T2^ = Vl^^^^iii^ =
sn it _ ^
= sn (^^ + 2g) _
cn(M + 2X-) ~ H^ - ^^n^*
A;2
sin=
+^
sin (^
tn (m
k^ sin^ e
'^g
(d)
vT^^^~F^i^'
X<t>+7r
22.
^^
"^o
and
dn
then
'^'^
'^
first integral
Letting
sin^
k^ sin2 ^
-'o
A;^
*" **
en u
tn?^.
Consider
en..
= VT^^^i^^F^ = VT^l^^i?^ =
d" (-)
^o
(a)
= Vnr^i?(Z^ . ^T^T^;^ =
f'^L^^^
VI -
Vl -
'^a
'^
Vrr^^r^,
dn. =
= ~
..
(6)
= - tn
tn (-.)
(.)
de
Probler21
,im
dn,.
j ^
^""^ ^^
""
From Problem
For
d
j- sn
(a)
(,
Jq
'o
k'
ypJT^k"' sin^
vT^^fc^
Prove that
de
K'
23.
[CHAP. 16
ELLIPTIC INTEGRALS
340
fe)
dv
- 1;^)!! - Ic'V')
/(I
We
ia)
shall write
By Problem
18,
dx
if
snu, then
A(snH)
du
du
Hence
(6)
in place of
sn- .
dx
dx
r
>'o.
(a)
V(l
from
to x,
dv
we
- a;=)(l - fcV)
=
sin
F{fe,*)
haye, since
sn-'a;
or
where ^ - am,
F(fc,
sin-
a;)
/(l-a;'')(l^^^fcV)
_
(corresponding to fc-0):
for trigonometric functions
Note the similarity with the result
r_Jv__ ^
^0
-^/l
The
gjj^-.^
trigonometric functions.
functions are generalizations of
elliptic
-u^
MISCELLANEOUS PROBLEMS
^_
24.
Prove that
F{y\l2,-rrl2.)
From
r^r(i)
2'V 2r(|)'
gamma
function,
de
sin~'" dg
Vsine
25.
2, this
Jo
V2F(Vl72.
integral equals
r(i)r(|)
\^r(i)
2r(J)
Tr(|)
./2)
at
and so the required result follows
of
Determine the period T of a simple pendulum
length
I.
attached
simple pendulum consists of a mass
mass and length I (see
to a rigid rod OP of negligible
suspended from a
Fig 16-1). Assuming the rod to be
equation of motion of mass
fixed point O, the differential
is
given by
rf2|0
^ _d^e
m?|| = -ms-sm* or
d'e _ dp _
de _
Letting
-^ - P' dt^ ' dt
.
dp
equation becomes
^2
= 4cos9 +
I
e.
g . ,
-ysmfl
dp dl
ds dt
-fsin
9.
__
dp
^ de'
the
Integrating,
Fig. 16-1
CHAP.
.
p
16]
ELLIPTIC INTEGRALS
makes an angle
iL?de/dt - r^t^"
when e =
then
e,
at
6 = >
= - (g/l) cos 9. We
d/dt
^^'^
^S/
fl
to 9
sign.
Vcos
time t =
thus have
Vcos
yligll
cos
and
released
is
from
rest,
i.e.
Then
^^
- V-^ r
^ "^90
341
-jz
=
COS
d.
Jf'"
2sr
-.//.no
fl
n= a
Thus
4^1
^^
=
_
2 ""
sin
26.
0,
Vsin^ 9/2
9j2
sin
sn{u
COS
+ v) =
sin''
(1
2 sin^ 9o/2)
becomes
^**
Vl -
fc
'
sin^
A;=
sin
eJ2
is
srmcnt;dnt + cn^sn^dnM
;
a constant.
a,
'''
2 sin^ #/2
9/2
4-\p p''
Let u + v
^
Vl -
Jo
2^/
ff
this integral
it,
T = 27rVW which
Prove that
4-\/ir
^
'
T
It
^'
Vcos
'^^
*"
r
"^o
k^ sn^
Then dv/du
sn^ v
-1.
Let us define
C7
snM,
y=
snt;.
It follows
that
^U
where dots denote
/.
= {l-U')a-k'U')
U = 2k'W -
(1)
(1)
by V,
we
(2)
V'
- V'){1 - k'V)
(1
find
{l
+ k'W
(2)
V =
2fe^y=
{l
+ k')V
2k'UV(U^ ~
V')'
(5)
^2y2
j/y
Dividing equations
(3)
and
(5)
jj2y2
C;y
=
(1
(1
- A:'Z7'V^)(y^ i7V+ UV
t/y- UV _ _
UV-UV
= ~{UV~UV)
_ fc2[;2y2)(y2 _
jjjj
^^^
7^)
(5)
we have
J7y
But
Then
and
UV - VV =
It
dV dv
U'
Multiplying
dV
and
j/y
2fc^t7y(;7y+;7y)
1
- k'WV^
(6)
d(UV- UV) ^
UV-UV
d(l
- k^U^V]
- k^U^V
'
[CHAP. 16
ELLIPTIC INTEGRALS
342
An
integration yields
c (a constant),
i.e.,
^^u^^^
sn
M cm; dm; + en m sn
1 fe^ sn^ u sn^ V
dn w
It is also clear
is
sn
-a;
u+v = a
that
M en f dn -u + en w sm; dn m
1
sn^ u sn^ v
__
y^^
_|_
is
a solution.
The two
^^
A:''
= 0, we see that
Then f{u + v) = sn (u 4 v)
Putting
/(m)
-y
sn u.
Supplementary Problems
ELLIPTIC INTEGRALS
27.
(a)
show that
to
<
if lx|
1,
Vl- = l"2~'i2J4
(6)
If
|fel
<
1,
Ans.
29.
30.
31.
Show
{a) E(yf2l2,rrl2),
(a)
E(l,<p)
Evaluate
f"'
Vsin' X
(6)
Ans.
Show
y/l
^f(~, 0)
that
3 sin^
(d)
in
F(l,4>)
ln(sec.;6
+ yVA =
terms of
in
terms of
^'
sm 9
f"' VI +
^o
(c) '(0.5),
(d)
sin"'
1.
tan 0)
[Hint:
sin
f^
Let
elliptic integrals.
elliptic integrals.
(v^
kV3/2).
2.1565
where
sin^ e
2 cos^ a
^'^
Express
-'0
33.
sin 0,
-y
VI
,r/2),
1.4675,
(c)
-'0
32.
F(V2/2,
(6)
(6) 1.8541,
(a) 1.3506,
that
V2-4-6y'8
{ -(!)- (K)"^(l^)"
=
Evaluate
&
prove that
J''7r/2
28.
\2-a)
t)
^2^
^C'^^""W|;^
(,r/4
3 sin 0,
j/
In tan
Ans.
.a/2).
2 cos ^
^K^-^J
W^^/
V^
be the para-
^^^-
^^^
ELLIPTIC INTEGRALS
343
f ~=J^=
Evaluate
34.
^""- X y^^jTj-
C -^^
--^{KVf.O-CVi,
36.
Show
37.
that
^k(
^ns
= J_x/^J_\
terms of
in
elliptic integrals.
()r^=^=
r\/3^ dx
re)
x'
'^^
(6)
38.
An.
40.
(a)
Show
39.
Show
(6)
X^(^/2,
46.
Evaluate
(a)
Prove that
^___^^^^____
Prove that
(a)
.MS.
47.
48.
(a) sn^,.
Find
Find
^ n(V2/2, -2/3,
1
L^V3
0.
k^
en
cn^,
cn^u
u),
(6)
(a)
(cn^M-sn^M)dnM,
(a)
-(tnu),
en
=
=
1,
1.
(5)
-^
Show
that
f^
50.
Prove that
VI
-A;
/^J
n^
(a)
dn^w
W
,.,
dn
u =
0,
IS
(a)
sn^xdx
sn(u
(e).mO =
0.
1.
- vT^T^.
A'
tn
(d)
1.
k' sn^
rf^
f"
en
2^
snwdnw
\r+lJi^
^;r;r~-
(dn )'
(6)
sech-
(6)
(.)
k- where
= l ' en 2m
l + dn2M'
(a) sn^zt
- (sn m
(6)
j en u du
sn
Ans.
w).
(a)
= icos-Mdn^.) +
^,
c,
(5)
(5)
={u~
+ K)=^,
dn'
=
f-^L
Sli
ty
49.
./2)
Ipf^
ELLIPTIC FUNCTIONS
Show that (a)snO =
45.
./2)
^^
r_^___ ^
z\
3
Vx* + a;^ + i
V 3 'sjx' + ^^ + i = (^' + x +
l)(x'-x + l).]
42.
44.
rfa;
J^^(|)
^'^
that
[Hint:
43.
(d)
a;
r^
that
Vs^cVi),
dx
(a)
(c)
41.
r"
it
In
en M
+ dnM
c.
E(k, amzt)}.
(6)
^"^
cn(u
="("
m-
^' ^""^
k'
+
,,.
+
-^)--^inr. (c) dn(M + i^)=.^
,
where
A:'
"
[CHAP. 16
ELLIPTIC INTEGRALS
344
51.
^
Prove
52.
Prove that
r_
,^ ,
that
dv
t^-^^^,k)
F(fc,tan-'x)
where
fc'
= ^/T^-
J_^ ^^___,^^^-^-^^^
k being understood.
briefly as tn-x, the modulus
tn-Mx,fc)
write
""
=
-
n en
2*'''
'
^^3"
Vx^
-1 Vs^
+ 1' 2
MISCELLANEOUS PROBLEMS
Prove that
53.
x^
sin-
dx
55.
Show that
any time
at
Problem 25
the angle e for the pendulum of
sin 9/2
where
56.
t is
r-r^^=e
Vsins + cose
-^
sin
cos
= V2
57.
58.
Verify equation
59.
Prove that
60.
(a)
cn(M
i;)
(6)
dn(M +
v)
Evaluate
Ans.
sin
eJ2 sn
(a)
(4)
62.
Prove that
if
en M e n
k.
(6)
dr^u
\v^
+ tt/A).]
sn m sn dn
fe' sn' u sn^ v
tt
dn
dnv - k^'snusnvcnucnv
1 - fe' sn'' u sn^
0.8044
{iS),
i;
F(V2/2,
^^^(^"="' ^)
V"/
,.,',
61.
is vertical.
of Problem 26.
_
-
(a) 1.1424,
sin (e
given by
{yfgllt)
pendulum rod
measured from the instant where the
Showthat
[Hint:
is
Page
332, using
^.
Landen's transformation.
then lim k.
1.
Hence verify
(15).
Page
333.
We
often
chapter 17
Functions of a
Complex Variable
FUNCTIONS
variable
z,
written
= /(.)
The
^rdaL^.,
^^^"'^'''^
"" 1'
o"
operations
^^ '^' '^^^^^'^
'^""^t
with complex numbers have
^.(.,.).
Example:
w =
z'
(x
+ iiiV
-^2
^.2
llZ
"=
/an^^^^^^^
^^^^
o-
given any
0,
there exists a
that
such that
/(^)-/(^o)|<e
o if lim f{z) = f{z,).
whenever
>
given anv
b- J<
A
J?
^Alternatively,
^V
"'
'
I
^ ^''''*' ^ ^
tv,
.'.
/(z)
>
continuous at
is
Z-*Zq
DERIVATIVES
by
f ?/),1f <;:tal'?t'""'
'"
" -'^'"
then
/(.) is called
anaiyL
/n
'
*- " <*<'*
Tnot
^^ ^
We define
e'
6,+
4,
From
^+^+
these
...
^i^,
we can show
345
^^-al
that
^
e'
z^
n
,
z^
= e-. =
and
extension of the
"^^ '""^^^^"^
^tH TeXeTbTr T^
^^^^^" ^'
?^
'T'"-'
'' -"^le-valued
L^^
necesSrty
correrp:nrr WtLr^f^i^l^^^^^^^^^
/(.) exist, we can use as definitTon Thrsells
Example
/W, denoted
(1)
In orTer to be Lalv^^^^^^^^
Of
AZ
As-^^O
continuous.
"
"'
'^'^"^^^^^^^^ f
z^
^ (eos
2/
sin
2/),
as well
346
Example
We
2:
much
variable are
Rules for differentiating functions of a complex
CAUCHY-RIEMANN EQUATIONS
A necessary condition that w =
is
du
dx
Problem
by differentiating
{2)
real
_ _dV
(g)
dx
u and
continuous,
v with respect to x and y exist and are
that
dx^
Thus the
dU
By
be analytic in %.
find
u{x,y) + iv{x,y)
equations
dV
By'
etc.
continuous in
If the partial derivatives in (2) are
7).
we
_
"
be analytic in a region
f{z)
(see
[CHAP. 17
"
'
dx'^dy'
harmomc
in
two dimensions.
Func-
functions.
INTEGRALS
If nz) is defined, single-valued
of f{z) along
some path
in a region
and continuous
zi to
point
where
z;,
z,
x^
+ lyu
z^
- x^^ iy2,
as
is
\(f(z)dz
where
is
an upper bound of
g
1/(^)1
f\f{z)\\dz\
on C,
i.e.
\f{z)\
M\ds
^ M, and L
=
is
ML
CAUCHY'S THEOREM
the region bounded by
be a simple closed curve. If f{z) is analytic within
that
as well as on C, then we have Cauchy's theorem
Let
j f{z)dz
f{z)dz
=
is
(5)
CHAP.
17]
Since
/(.)
2.
f{z)dz
has a
analytic everywhere,
is
(5) is
IiL^\nteiritvelpei1n"?ht^^^^^^^
Example:
347
2zdz
i-
2zdz
we have
11
(1
+ iV -
i2iy
2i
'
''
to
c!\hel
/(a)
where C
'P^'
1
"^"^^^ ^"^^^
^ ^"d
fi(^d,
s.-f r
'"""
n''
rrvr?thrst'i\Tow
'''
TAYLOR'S SERIES
Let f{z) be analytic inside and on a
circle having its center at
points . in the circle we have the
Taylor series representatrn of
z-n
/(.)
/(a)
+ f'ianz-a) +
T>,nr,
fr.
11
'"
Jiven by
^(.-a)^ + g^(.-a)3 +
..
(,)
SINGULAR POINTS
Example:
If /C^)
^^^,
then
is
an isolated singularity of
f{z).
POLES
^f
irntk
^(^)
= (^:r^'
<^()
^ 0.
where
:x; o:rrsv-iiL
doz.6?. pole,
etc.
'
poTe
</,(^)
is
fitt'^
i,r^^^'
''^^'" '^""^
^ ^^"^^^^
'
n=2
it is
called a
Example
1:
/(.)
(^^ife+l)
3^~1 =
^i^TJ -
f(,\ /w
or double pole at
^^^ *" singularities: a pole of order 2
Zz-"^
(^
= 0.
(see
+ 2i)(z - 2i)
^ simple poles at z
Problem
37).
The function
Im^
3.
= -1.
or simple pole at z
[CHAP. 17
348
is finite,
2i.
For example,
/(.)
= ^IM
we
call
= V^
/(.)
has a singularity
such a singularity a
removable singularity.
LAURENT'S SERIES
a
every other pomt mside and
has a pole of order n at . = a but is analytic at
all points mside and on C
with center at a, then {z-arf{z) is analytic at
on a circle
and has a Taylor series about z = a so that
If
mC
H.\
/()
^-"
(2_a)
4-
^-" + ^
(z-a)"-i
+
^
+ -^^^ +
z-a
...
ao
,,
+ ai(z-a) + a2{z-af +
...
{9)
ij ^alj^d
a^{z-a) + a.iz-af +
a Laurent series for fiz). The part ao +
the
consisting of inverse powers of z-a is called
the analytic part, while the remainder
refer to the series
t^f^i^-af as a Laurent series
This
is called
principal part.
More
generally,
we
A
principal part.
where the terms with fe<0 constitute the
center
having
a region bounded by two concentric circles
panded into such a Laurent series (see Problem 92).
function which
at .
=.
analytic
is
^5^
uZ
sometimes a pole of
Example:
infinite order.
The function
e^"
+ | + af? +
'
'
"
^^'
^'^
RESIDUES
^'"^"^'^^ singularity at z
= 0.
^,
^^
The
crfficient
It
.^.
coefficients in {9)
a-u
where n
is
of
is
a_i
RESIDUE THEOREM
lim(z-a)/(z)
(^^)
is any
except for a pole of order n at z = a and if C
If fiz) is analytic in a region
Integrating
(9),
form
the
[9).
a, then f{z) has
containing z
simple closed curve in
^^^^-
1'^]
dz
<r
Jc(z-a)'^
Problem
(see
13), it
"""'' '
^Lm^^S afplfe!
More
grenerally,
Theorem.
fimte
we have
*'"' "'"
call
27rm-i
(_^^)
>*
"' ^>
number
of poles a,l>,e,...
dz
/(z)
i.e.
(12)
Cauchy
is
27ri(a-i
sum
6_i
c-i
of
-5?
excent at a
respectively!
(_^_4)
EVALUATION
^ "-
.)
"
J>^
by C.
ifn^l
ifn =
follows that
^^f{z)<iz
the
JO
\27ri
349
this result
which we
DEFINITE INTEGRALS
F{x) dx,
F{x)
Consider
G(sin
e,
cos
Let .
^l^
See Problems
e) de,
..
The given
dz/iz.
along a contour
X ^(^)
Here we consider
Type
4.
1.
is
'-^,
and cos 6.
dz
where C
is
cose^'-+f^ and
integral is equivalent to
29, 30.
^me =
Then
^'
^"^^
tZ!
ihen 7f
let i2-*oo.
^
an even function.
F{z) dz
"^^
p"
J-2ir
is
F{z)dz
ie-d0 or de
^(*)
a rational function.
F(2)e-dz
where C
is
[CHAP. 17
VARIABLE
FUNCTIONS OF A COMPLEX
350
Solved Problems
FUNCTIONS, LIMITS, CONTINUITY
Determine the locus represented
1.
21-3
<n\W
Method
(a)
12-21
(b)
\z-2\
1:
with center at
\x
c^
^q
1".
3|
lz
+ iy-A^\^-^ + ^y
and radius
(2, 0)
,.,,,
z-31 + +
^^^
+ 4,
12
3.
+ Oi
and 2
If this
Squaring,
Method
(&)
1:
la;
-1, a straight
line,
line.
any pou^t^on
that the distances from
Method 2: The locus is such
perpendicular
Thus the locus is the
to
,
())
(2, o)
and (-4,
^^
l^^^nX/^^S^ir^^'
T'^i^^m^:''
q
10- V(^ + 3) +1/
_
x^
2/^
25
16
0)
find
are equal.
^^ ^
J^^
^ _^
^^^W^V
^^^L^: l^^'J^:^^
(c)
we
+ t2/-'i|-l'^^2/-ri|
squaring
+ ^^ - bVI^TbFT/.
of
semi-mmor axes
with semi-major and
ellipse
an
'
on
distances from any point
to (3,0)
it
and
2.
^""^
'''
(b)
1<
+ 2i| g
|Z
|.
by each
represented
in the . plane
1.
2.
2i\ is
^v, +
so that
the distance from . to -2^.
J+ 2^=-
\z
of the following.
^^\
1
1
+ 2^^"
I, 4- 9,-l
bu^SIr
ow
1
1
is
center
a circle of radius 1 with
-^'^^^.^
1^^^^^^
^^en
=
1^ + 2^1
2.
<
1.
2^|
below.
(C)
7r/3
arg
^/2.
z-pe^.
Note that arg^=*, where
g^Fig.
Hg.
these Imes. See
mcludmg ^^^^JX^
=
W2,
=
and
^/3
by the lines ^
^^^^
below,
17-l(c)
1
J^'^-"-^
Fig.17.1 (b)
<"^
(b)
_s
()
1/il-z),
(c)
the form
e^
u{x,y)
iv{x,y),
{d)lnz.
(i.)' =
+ ,,)3 = .3 + 3x^(..) + 3x(i.)^ +
i(3x'2/ - J/')
x' - Sicj/"
=
x'-3x2/S v{x,y) = Sx^y - y^
u(x,2/)
Then
.
=
(x
-I-
x'
+ S.xV-3x.-^r
real:
CHAP.
17]
w = -J_ =
1-z
(6)
T'^.-
(c)
Inz
(d)
Inp
Prove
We
i^
sin (x
(6)
cos (x
-f-
+ iy)
sin (
isr.Zy)
(J3^^.
tan"'
Yi
sin x cosh
e-
sin
a;
cosh y
cos ^
Similarly, COS.
e^^
sinSy
tan-'y/a;
2/
+ i cos x sinh 1/
- z sin sinh y.
a;
sin
T^
e- =
z,
^1^,
'""""
2/
cosZy,
cos x cosh
+ %)
^""(cos ^
e^^
and
j//a;
+
+
cob (x
sin
a;)
+ iy) =
sin
z,
from which
e'-'
e-'- +
2i
e^cos
cos x sinh
= !!+^Z!
COS.
e--^""
cos ^
2i
=
=
and
+ y"),
ilnix^
=
=
sin.
vix,y)
1-x + iy
- xY + y^
(1
a multiple-valued function
is
iy)
1-^ + % _
1-x + iy "
V^?T?'+
In
'j;^^r
t-is
^ < 2^ and
(a)
Then sin.
'^^
vaL: lor
vame
foVwhTh
wnich
4.
l~x~iy
e^^(,o.Zy
u
Note that In^
= O-^^Ty^,
^3. ,3.
ln(pe'*)
-(-.-)
,3...,
} + ty)
l-{x
35^
e-"'-^">
a;
sin x)}
(sin
a;)
(j~^)
e(cosa;-isina;)}
i(cos x)
('^llHZ')
j/
e+"'
i{e-
(cos X)
(^^
= ^{e-Mcosx +
e-''^^+}
y ~
2^
^(s'n
')
(^'"2^
isinx)
'
) =
definition.
which A.
= Ax + iA2/
=
lim ^
tyZl
Ay
0,
=.
j^^/(Mz^^
approaches zero.
lim ^
y-z
If
A/(,)
^^
^V
is
Ax
0,
is
lim
lim
Ai(-o
th^t'^mZV^T^
isinx sinh 2/
x-|-%-f Ax
ji
Ax-*o
If
Then
Ax - i Aj/ ^X + iAy
cos X cosh y
(a;
- %)
tAy
Ax
^,^0 Ax
=
Ax
^^
=
lAy
iy
iAy
i
A!/
-1
A.-O
so
m = j^,
U w=
(a)
6.
j^
Method
(a)
1 +
1 Hm
-^
1:
(1
Method
2:
The usual
^.
find
+ As)
+ Az)
(z
(z
(6)
+z
1-z
/(z
(1
- z - As){l - )
-^ 0.
z^l.
w =
f{x
+ Az)
= 1, where
is
,.
io
u{x,v)
dv
du
f(z)
equations
a region is that the Cauchy-Riemann
the region.
/(z)
_
~
Since
non-analytic.
^ 1,
The function
is
Si
provided z
zY
Determine where
{h)
differentiation,
[CHAP. 17
352
g^
e^
+ iv{x,y) to
dv
du _
- "'d^
e^
iv{x
the
be analytic in
aatisfied in
^ satisnea
'
i.e.
+ Ax,y + Ay)
Then
f(g
+ i^)-fM
1^0If Aj/
= 0,
_
-
Ai
u(x
,.
u{x,y)
^^
+ i{v{x + Ax .v + Ay)-v{x,v)}
+ i^y
is
u(x + Ax,y) -
,.
+ Ax.v + Al/) -
1^0
fv{x
u{x,y)
+ Ax,y) -
du
v(x,y)
If Aa;
= 0,
is
we must have
.av
dx
dx
dv
du
- ^
and
+ Ay) -
v(x,y) \
13m
i
3t
^^
9u
dv
aj/
_^du
av
dy
dy
ia^/
^v
att
{ v(x.
so that
ai;
dx
dX
dy
_^
i.e.,
Bv
dy
du
g^-
8.
(a) If
by using
(a)
m^uix,y) + iv(x,y)
f(z)
is
z^.
u{x,y)
Since
(aJo,j/o).
du
Also since
= u^dx +
dv
Uy
Vxdx
dy
0,
v,dy
we have
_
-
dy
^
_
-
Ux
Mj'
_Vx^
^^^
= uo,
v{x,y)
= vo
which intersect
CHAP.
17]
353
= Vy,
Uy
~^
so that any two members of the respective families are orthogonal, and
thus
(6)
f{z)
The
a;2-j/2
Pig.
then
z",
graphs
of
Ci,
u = x^-y\v = 2xy.
members of
several
2xy
= C2
are shown in
17-2.
Fig. 17-2
9.
m=
(a)
By
U)
Method
1:
Method
where
2.
Integrating
(i),
r^,
2xy
(2),
2xy
''^'^
""'"
_ _
d^
a^-
dx~
Then
= 22/-2
Tx
+ Ay + F(x).
- 2x + G(y).
+ "'
""^'^
'''-''
-^^^^
"
- -y --1 --t-t-
(a),
Ci is
f(z)
=
-
= ^ + 4^ + ^W- Then
^
_ ~2xf+ c. Hence ^ = 2a;2/
^ + i^
(x'-y'
%-
(2),
2y
+ F-(x) =2y~2
Az
2iz
c,
=^
+ iy,
= x-iy
so that
z^dz
i+t
(c)
We
+ 4i.
have
dy)
(1,1)
''(2,4)
X(2,4)
j/=)
dx
'
2a;j/ dj/
J"
"
2a;j/ <a;
(sc^
= ^^
drop out.
(6)
or
2a; -h c.
The
2i
substituting in
-I-
(a)
Thus
2:
F'(x)
r
KX) i"
Z.
and i- (a)
~ -fand'^iJi
From
Integrating
T=lxyV-t:M
(6)
- ^
^
dx
dy'
- j,') dy
[CHAP. 17
354
Method 1:
The points
(a)
become
and
(1, 1)
(2, 4)
correspond to
=1
and
=2
respectively.
,.
..
i
line integrals
C
t
(5)
{2(mt')dt
(1,1)
and
(2,4)
2/
{t^-t'){2t)dt}
t
"T"^'
-t
- 1 = aZTJ
"
(^^
^^
1)
1/
3x
- 2.
Then
find
{[x^
(3a;
- 2)^]
*^j=i
From
+i
to 2
+i
+t
to 2
+ 4i
C
(|+
Adding,
Method
2)3 dx]
(x^-l)dx
dy
l,
+ (-30 -
9J)
= 2,
dx
-f
6i
-g
3i
and we have
(-i-y'Jdy
at
-^ -
and we have
2xdx
-4ydy
3i)
2x(3x
i\
r^
From
dx
r^
+
<c)
ij
2(t)(n2tdt}
Theline joining
we
{{t^-t')dt
30
9i
6i.
2:
By
11. (a)
(b)
f{z)
dz
then
"
"
^ ''^
0.
^^
J^
V(z) dz
is
Pi and P2.
^/(^)d^
(a)
By
.^-,.,
where
is
(b)
Then
jj(--it)-*.
^c-^'^"*
//(s-Sh*
Since /(.)
zero.
^ {u + iv){dx + idy)
is
f = g- = "i^
analytic,
^ /(z) dz =
0,
assuming
/'()
^""^ ^^ *^"
(P-^^^l^"^ ^)'
r
r
Then
mdz
/(z)dz
f{z)dz
+
-
=
/(z)d2
/(z)dz
/(z)dz
Fig. 17-3
PjAPj
i.e.
P.BP. (path
f(z)
z"
is
2),
analytic.
is
CHAP.
17]
355
12. If f(z) IS
As
dz
/(z)
X^
az
f(z)
AB
cut)
a cross-
(called
f(z)dz
cr By
AQPABRSTBA
Fig. 17-4
md.
But
f(z)dz
An
^"
^(,)^^
Hence
(^)
U)
^^
gives
dz
Then
r^^^^^
BRSTB
= -jf(z)dz.
//(.)
J/(,)^,
AQPA
and
/(^)
dz
f(z)
dz
I.e.
^
need not be analytic lithin
curve C
Note that
13. (a)
f(z)
f ^L_
Jc{z-aY
Prove that
= j^^i
if
if
What
(a)
is
=.
n =
n =
Jc
(^
- a)-
'
where
3 4
2,
if
having center at
that on
Fig. 17-5
e" e'
w-1,
pll-nHe
"Ode
^i:
the integral equals
P'
rf
14.
Evaluate
(a)
\o)
ifn^l
.here C
is
is
not intprinv
"'^''"''^ tn
t Ul
l^i
Since z
is
interior tn
mterior
to
Smce
""
2-j
\Zir
e-Ml^^^L
^0
(6)
a simple closed
y^ J^^r^
If
is
=a
0,
To evaluate
f(z)dz
a as interior point
/(2)rf2
J^
U
|. +
-i- i
^j
(a)
- 1.
_
^
_ 4,
1
the circle
,.,
!.
--ywhere
""
is
(,)
zero!'
the circle
,...,
'
al
11).
inside
[CHAP. 17
356
closed curve C,
analytic inside and on a simple
is
15 If /(.)
is
, V
integral becomes
Letting z-a^.e^o, the last
continuous. Hence
and .
is
we have
f'" /(<>-+
Ha, +
J^
d6.
ee'*^).
ee
But
i)de
_
lim
it
2)ri/{a.)
Ja
-1)
^
.
Evaluate
,
16.
(a)
Since z
^^o^^rfz
az,
/^
(a)
,.
-r
lies
Then
ir
... ^
withm C,
where C
^_-!l-.d2
J^ ^^^ ^ y)
(M
^oj
j)^ ^^r:;;
J_<^2ilcte
"*
2^:^
<C^^d. =
J^
-^
COS.
is
the circle
j^-lj-S.
by Problem 15 with
= -1
/(z)
cos
-2.i.
'c
by Problem
17.
IB, since .
2s-ie
and
Evaluate
^^^^^
Method
1:
By Cauchy's
n=
If
and
/(z)
13.
= -1
= 5Z-3Z + 2,
lOjri
2>rie-'
2,ri(l
-e
')
^^^^ ^
integral formula,
=
by Problem
^^
f"Ma)
then f'(l)
10.
2^ J^ (z - a)"-"'
Hence
'"'""'
'
''
z,
CHAP.
17]
357
18.
^"^
'^^'^'thterni
-?>;?2^-
converge?
= -^.
2"
Then
M^
lim
By
|i^!Lij
lim
|__L!^_^
converts
'verges
test fails.
\z\<9
n^ diverges
a< 2 and
if
it
(6)
2.
2
^'
(2m -1)!
3!
pl
We
5!
(-1)V"-^'
(2m + 1)
lim
Then the
series,
2^^^
3"
We
The
(2n-l)!
'
have
Mn+1
lim
lim
X2"
converges
\z\
if
Locatejn the
i\
<
3,
lim
all
values of
(z-^^)^'+'
Z
^
"""
2z'~z +
|2m(2m
'-*
and diverges
|a|
R-
,
'V
anz"
M.
sin Tnz
^
+ 2z
+ 2
,
since
circle
'""'^
i.-i,
..
z-i
3"
\z-i\
if
>
+ 1)|
- i)"
3.
^'"
''>'
"-
can write
.^
= ~^
1
~
show that
uniformly convergent
it is
.^^
is a
a ti,^Iq
'S
pole
it
[^l
M.
converges,
it
follows
/J.
^f
J
o
of order
3.
m v^ 0.
+
^^^verges since
Since by hypothesis
^-mz~i)W^T+2i)
rr
z'
= 2,
\az
M test that
thnf
finite
7T~r~7T7!)'
,^s
\z\
the ratio
and on the
is
;^
and
'-">
=2
z.
we have
by the Weirstrass
(6)
is
In this case
\z
If
(-!)"-> z2-l
-tl^lrmllstJ'aTproIcl^erras'MV!:!
20.
2.
have
series converges if
Thus the
19- If
>
-.
^
oil points
at all
inside
i p
i-e-
{z
the
\z\
(~l)"'''g'""
lim
(c)
<
=
Izl
1^1
if
?, ^HFZ^
converges.
Thus the
k(
|-l.
5= in^2"
\z\
-j.
\z\
w''2"|
2.
+2
+ 22 + 2 =
. {. - (-1 + ,)h. - (-1
Since
^'
poles:
when
-,)}
z
= ~2
(,+
= ~1 + i
v/4
- 8 ^ -2 2i _
_ ,)(/^ , ^ ./
and z
= -1
-i.
2"
- 2i
are
We
-1
i,
[CHAP. 17
358
1
(d)
ljl-21i.
^
*
cos z
lim
However, since
appears to be a singularity.
0,
is
it
removable singularity.
Another method:
^=,
2!
6-'--"^
This
Then
is
1 is
an essential singularity.
finite
,.
leUin.
1^':^^^^'^^^^^:^
of
nature of a possible sn^gularity
in general, to determine the
u-0.
at
function
new
of the
and then examine the behavior
of radius
points inside and on a circle
If /(.) is analytic at all
that^
theorem
Taylor's
^is any point inside C, prove
see that z
terms.
infinite number of non-zero
the principal part has an
a Laurent series where
essential singularity at
has an
Jf
we
2U^^^~""
^"TT^^
21
'
4!
a removable singularity.
is
(e)
U =,-+
z*
4z'
if
f
= |{l-(^l-2r+4T--6T+-V/
^^^
1
Since
/(z) at z
By Cauchy's
+ /.)
/(a)
/./'(a)
+ ^/"(a) + ^/'"(a) +
_ 1/.
.,
and
jC
f(z)
dz
(i)
substituting
"(F=^)[l-'i/(z-'^
(.) in (i)
formulas, we have
and using Cauchy's integral
on ^
C,
the length of C,
,
,
when
Now
n-
=0
IB J
z is
-. 0.
^
|j3^-rx| /(^)
Then R.
2^i
and
"""
*"
...
^<f7t^r^
a)+Mz
J<, (z
\z-a\=
"
B,
?i)
so that
since
by U), Page 346, we have,
'
l;,|+iM
r,
be obtained by use of
.^ r____M^z___^
-a-
_
"
D
^"
m^
r/(^)d^
where
As
let z
division.
i-a-h
2fffl is
.e
we have
1
By
with center at
/(a
known Taylor
series.
.<
u_i
iS
r.
we can
CHAP.
17]
359
Find Laurent
22.
Name
e
(")
(2
_ 1)2
Let z
1-
(z-ir
-1 =
The
(6)
2 cos -
''^>
sinz
sin
z
7^
^''^
fz
Let
rfowftfe 'pole.
n-
Then
M.
+ ^)
sin (m
+m
is
0.
8z
z-0
._
is
and
^l/
_
~
it'
m=
31"^
5T~
7\
-1-
+ 2)
z.
Let
i+
a pole of order
1,
1)
- 2m +
2
2(z
16
16^
a poie of order
32'
1,
2m^
2m'
+ l) +
2(z
..
+ ir
or simple pole.
The
,r
_ sin m
-7+1 +
i,
-- +
Case
1.
L.^..^
(z+l)(z
The
a removable singularity.
is
JT
+ l^rz + av
sinz
The
+ ~4!
3!
f!+fr+-
an essential singularity.
is
_
- T.
or
and
i?|i + + |7 +
^-l ^2! ^
- -i- + ^_.
^-'^
+u
= 0.
The
2,
zcosi
z
a poZe of order
is
"^
lir
(z-ir
z
Then
M.
or simple pole.
<
|z|
<
2.
<
|z
+ 1| <
1.
[CHAP. 17
360
Case
2,
_
"
^- -^-^
~2m'
1
^
+ if
z{z
= -2
23. If
'
is
which
is
Let
+2 =
4m^
4(z
2(2+2)'
a poie of order
and the
+ 2r
where a-n
0,
(z
<
<
2|
_ a)" +
(z-a)"-!
prove that
(a)
By
integration,
f (i^'^^
27ria-i
Multiplication by
(z
- a)"
13
S'^
Taking the
- l)st
o-
call
f^{ao
a-i(z-a)"-'
lim
z-ta
f{z)
at the pole .
= a.
we
find
dr
^ ^ )" ^^^^^
;S^
dz"
result follows.
indicated
Determine the residues of each function at the
+ l)'
+ a.(.-a) + -(.-ar4-...)d.
+ a-n^Az-a) +
(w-l)!a-i
from which the required
we
(z-arf(z)
(z-2)(z^
2-a
27ria-i
(a)
+ 2)
32
2.
remains,
Since only the term involving a-,
24.
1^(^
RESIDUE THEOREM
/(2) dz =
(6)
3.
(a)
^m^'
16
+ 2)
8(z
/(2)
-M-t<ty^(i)'<f)'-}
-2m'(1-m/2)
L_J^_^--i
16
8m
32
is
Then
u.
(u-2)v?
The
RESIDUES
-2.
2,i,-i.
Residue at
is
lim
Residue at
=i
is
lim
Residue at
= -i
(z
Then:
4
2)
(z
poles.
poles.
(^
_ 2){z'' + 1)
5'
1
v"
- i)
(i
- 2)(2i)
r'
is
J.
(^
+ i)
i (z
- 2)(z - i){z + i)
(-i-2)(-2t)
- 2i
10
1
+ 2i
10
CHAP.
17]
^^^
z(z
'
+ 2)
'^
0,
-2.
Residue at 2
Residue at .
= -2
lim z
is
;_im^
is
= -2
a simple pole,
is
a pole of order
Then:
3.
is
361
^ ^^ |(, + ^f
^^j
_i
2A ^
8'
Note that these residues can also be obtained from the coefficients
of 1/z and l/{z
the respective Laurent series [see Problem 22(e)].
ze'*
('''
ig
_ 3)2
'
3,
Residueis
lim
^ |(, _
3).
_^j
Then:
^ limf(.0
=
cotz;
(d)
a pole of order
5s-,
Residueis
a,b,c,
lim (e-
Ste
(-1)(-1)
= (
lim cosi)
f{z)
dz
{sum of residues of
2,71
reasoning
similar
that
to
f(z)dz
For pole
lim
except at a
number
f(z)dz
Problem 12
of
f(z)dz
...
a,
Fig. 17-6
J?/
^^
'
b.
O'
"
<h
f(z)
{z-a)"
f{z)dz
'
fit
(T^
ao
ai(z-a)
...
2rta-i.
^^ +
^
so that
...
/(^)rf^
^-^
=,
5,
6^(,_j)
...
27rib-i
f(z)dz
-^-\-l)
<k
.<-)
interior to C,
J^
By
rule,
variable.
i^ Y
( Um
=
Where we have used L'Hospital's
e"
Then:
1.
(z-5.)-^ =
Urn
+ 2)
2rt(a-i+ 6-1+
...)
of poles
362
26.
[CHAP. 17
""^" ^
r^^^TW
/c(2-l)0
Evaluate
- -3
lim
1 is
(z
-I
^^^^" ''
^^
1)
(^T:1)(z
Since
Izl
'''
+ 3)'
''''
'''
^'^
'''
16
is
-5e-
(z-l)e'-e'
7
,.
+
^"'sd^i^' ^^'lJ^iH^+sy
ia)
"^"^
lim
-^^
^--3
(z
r^2
1)
16
- 1,
z
3/2 encloses only the pole
TTte
Since
(b)
\z\
and
EVALUATION
27. If
constants,
are
where r
is
\16/
= -3,
2n-il
^^ J
^^g
DEFINITE INTEGRALS
of
1/(2)1
^'^'^
for 2
where fc>
Re^',
and
\\mymdz
prove that
22
=
shown
in Fig. 17-7.
By
j
f{z)dz
L =
that for z
Show
If
U/r
so that
M=
2,
_5^
dz
f{z)
Re'o,
|/(z)|
Re',
for example)
lim
ttB
\m\
=^'
^>'^
fe
+ R*e*'
= 4.
lim
and so
R_>
f^'''>
^^
CO
f(z)dz
|i?*e^'
inequality
Note that we have made use of the
1+^
1 ~ R*-l
\zi
+ z,\ g
*/p
<
|i
Fig. 17-7
R"-'
Then
ttB.
R-.00
ss
\f{z)\\dz\
28-
we have
" R*
\z^\
if i? is
with
zi
B*e*'9
and
R>2,
zz
1.
dx
29.
Evaluate
^
^0
Consider
_B
to
-^
f -^
-W C
and theCtcircle
Since
z^
+l =
T,
when
is
sense.
traversed in the positive (counterclockwise)
z
C.
poles e-"* and e^"^ lie within
Then
D-
Only the
CHAP.
17]
Residue at
e'^"*
Uz -
lim
+
~
e'^*^*)
j-eiri/4
Z*
1^
Residue at
e=""*'*
-ij
lim
lim \(z^^,3itiny
lJ
^e'^""'-
e^^'*)
'z*
+ ij
Ig-^"*
-ij
lim
363
Thus
(1)
f" __^_
Taking the limit of both
*Tl
30.
Show that
The poles of
and
= 1 + 1
(^2
=i
S<Tl'
lim
is
= -1 + i
dz
X>f
r
31.
dx_
we have
- i)'
of Problem 27 are
of order 2
(z
+ i)\z - ly (z^ + 2z + 2) J
+ 1 - i)
+ DVz^ + 2z + 2^
+ ir(z+i-i)(z + i + i)
2^/^1^^+3-42!
^
,
+
^
"^
100
^''^^'^^
+ l)V + 2a; + 2)
100
5!
('
^'^^
(z^
= 9i-12
?!
(z
lim
is
J_^(a;='
^(a:^
we
28,
J (
2--l
Then
yyfi
^'^^
dz
1.
--
Residue at z
lim
^ 1)2(^2 -(-22 + 2)
of order
Residue at z
,,,ff^
4.
~"^K~
_
-
^5~
7ff
Kn
z^dz
Xf
3-4i
_
-
7;
Jr(z^+lf(z^ + 2z +
50
2)
Evaluate
Letz =
e'.
^^
3 sin
Then
sine
3sine
^1^ ^ -^
dz
X^r_i_q,'*
A-^-A
i6
where C
is
ie'o
do
izde
so that
2dz
.7:3z^ + lOZ^
.^-~,
[CHAP.
364
The
1 he
'^.
of
poles
Doles oi
^^^ *^ ^^^^^ P^
_ s3
g^2 4. 10 jz
17
+ 36
-10i V-lOO
=
Only
Fig. 17-8
Residue at -i/3
Then
32'
,
Show that
If z
Jo
e<9,
Um
^
cos
de
+ -'
2
12"
.o
'=^3''
'
_
-
e"
e-""
(z
cos 3g
4 cos
Residue at z
is
Residue at
=|
is
+ z-')/2
d2
iz
pole of order 3 at z
+1
- l)(z - 2)
z(2z
^ i^
for ^
JJm
-j
(2
- i)
^^
^.s^gg
the semi-circular
65
24"
i/2^,j21_65l
-2^{2^t)^
fc
>
and
J e'-^'Mdz
arHf
JL
as required.
12
24[
Re, where
within C.
+l
- l)(z - 2)
dz
=i
21
Hm
is
z+l
1 d'
io2!di^r *z(2z-l)(z-2)
where r
izAe.
z + l
1
" 2i X7'(22- l)(z - 2)
Then
|;(,),
d;,
33_ jf
_ zl+^
2
is
rule.
^,
2i
where C
by L'Hospital's
4i
''
4 cose
'^
Then
V4V
cos 3e
5
- 3 / ~ -3 6 + lOi
3 j\__3z2 + lOiz
{^2
^ 2^/^
+ 10t2 - 3
2^^
32.
-3i, -i/3.
are constants,
prove that
constant.
Ifz =
B.',
J^e-/(2)d2
J^--''/(i^e')^i^e*<'d9.
d9
Then
ja/O
'
Jo
g-mRsinfl |/(Be*e)|
iJdO
'0
M
"= ~
Jn
"
is
a positive
CHAP.
17]
Mow sm e a
Now
28/-n-
"^
than or equal to
for
^"'^
<
=
<
=
"
2M
/o
-/2
(see
ti
365
Problem
Chapter
77,
4).
Then the
f"
R'-' Jo
As
Show
34.
ie -
this
00
r"2i^
dx _
that
Consider
f^^dz
=i
lim
is
j)
(z
Then
2*
and
~ i){z + i)
=
dz
r"
,-
rfj.
=i H
lies
i, but only .
'^"^'^
I.e.
fe
within C.
2i7i
sin ma;
(~
\2i
r
p.m*
so
35.
we
Show
i?
and usin?
Pro>,lo,y,
that
r^l^ dx
^K-n
i.
around r approaches
^ ^
we cannot
we modify
T^Z
''^;"''*!"
abde%gh1
Since 2
or
proved.
2i
dz
zero,
IS
i
is
poles at z
<|(z
and
,-
where
^he
path at
"'"' "^
is
outside
"'''
0,
inte-
that
as shown
^^"*-
^'
C, we have
~dz =
Fig. 17-9
dz
BOEFC
Replacing . by
-.
-da;
-I-
^''*
or
2i
r^
dx
^ d^
dz
-f.
BDEFG
EFG
dz
BDEFG
integral,
we And
[CHAP. 17
366
first
zero.
on the right approaches
Problem
Let .-^0 and R-^^. By
approaches
integral on the right
"'
de
The
^1
Then we have
r" sin X
lim
2-i
or
VI
dx
. ..o
equations. Then pomt A(2, 1)
as the transformation
B
pomts
Similarly,
Since = ^^ we have u = ^^^f" 2/
T, 7,^1 t>
figures below).
rJane
plane (see fig
into point A'(3, 4) of the
j ^^,^^^^^
in the x plane maps
-^P-^ivdy. The hne
a,^,^' with equaand C map into points B' and
;^
A C B C A ii ol cu
segvents
parabolic
map respectively into
17-10(a) and (6).
tions as shown in Figures
.
-2x
"
.fl
ABC
^^
-^-
C'(T,24)
(2,1)
1/
(4,3)
Fig. 17-10
Fig. 17-10
(ft)
The
4(1
curve
slope of the tangent to the
u^
curve
slope of the tangent to the
The
i-^
at
2v
3-i
m,2 irii
4-
OTim.
(3, 4) is
at
^^
^^ -
is
1,
RT'
BC
is
,s
21
d^|<3_,,
*i|
^1
^U
= M =
2'
3.
^^|^^^^
by
given
and
dv\
w.
(3, 4) is
17/4
(3)(i)
^ 4 T" oT,H
i_
i.
and
between
we can show
function.
+ w.)
tan
Similarly
(6)
(a)
AC
between
7!-/4. while the angle
./4
^^ ^^^ ^^^_
is
Jj^^/^^J^, T, a transformation where /(.)
^^^^''^f^t^.^/ing at . == .0 has the same magnitude
CHAP.
17]
FUNCTIONS OF A COMPLEX
VARIABLE
-=1
.2
Then
e<9_
[corresponding to
w = -i/j = gie/a
A and i- in
"
367
"'"'
-t"-!
to
tart-
Fi^ ITTll^d^)].
*'
*'
'^^''""^ P''^'"-
^hen
and
plane
w plane
<X
Show
S.
f "^'^^
^
that
Jo
Consider
rn
xJi^tT^J
^
= i =
pole z
g-^t
lying within C.
jg
Then
XT+I"^^
or,
= ~l
2e<>-'r.
2jrie<!'-i)'H
^B
We
BDEFG
CH
HJA
thus have
2.
5Z>Sg
'
'^''
^^
"'"'^^
^*
"^
^^'
J^.
^"^^^^^
^''''^^ ^^^^
-^ent
of .
=
is
''^^^''"""
increased by
[CHAP. 17
VARIABLE
FUNCTIONS OF A COMPLEX
368
'""""""'
-,.
2,."
or
so that
1"^
Jo
Supplementary Problems
FUNCTIONS. LIMITS, CONTINUITY
by
Describe the locus represented
case.
each
in
figure
a
Construct
39.
+ 2 - 3^| =
(a) |.
o o^
,
+ ^Y + {y-SY = 25, center (-2 3),
- 2^ + 1/' = 4, center (2,0), radius
Circle (x
= 1. where
Branch of hyperbola a.V9 1/V16
,
Ans.
(c)
Ans.
41
(b)
(c)
Interior of ellipse
Ans.
(6)
(c)
(b)./(3
^ln((l
,,,.
(a) lim
+ x)^ +
.
_
-
Prove that
.3.
,.,
(6)
Show that
(c)
..
z''
.n.
(M
C)
,.
Zo,
42.
^^ (^
w -
(a)
T^
If
(b)
For what
Ans.
(a)
1^
- ^1 =
+ +
CO
0.^
>'
3.
3|
x+y
;^;^;"7;7
10.
1. + 31 <
^^
)_
^
the
9,
f^.^
/(2)
z
^
finite
l-l/2^
+
^
find
- 2/_^+
a;^
+ 6x +
2/^
and
values of
(ft)
z is
^^^^
^.^^ ^
^^
/(,)
/^^'
a;
"
+f
3x'^
+ "*-
(a)
are real.
-(^..). where . and .
= tan-rf^+2/..,
=
is
/c
continuous at .
a,
1,
0,l,2,...
P.0,. .h..
0.
(6) to
the equation z
1.
EQUATIONS
DERIVATIVES, CAUCHY-RIEMANN
44.
+ 51 "
1.
2.
""
(^. =
form .(...)
+ .), (c)e=%
=
x^-Zxt-^y, =
M
a;^ + 3x + V^
M = ^. + 6 + 1/^ + 9'
=
M = e^'-"' cos 2x1/, V
+ 2i.,
(a)
+ J/V16 -
V25
(,)
quadrant bounded by
in the
1^
2.
- 2) + ^ +
(x
circle
(a.)
(")
radius 5
-^ *
plane represented by
(5)W^3. Ogarg.Sl-
1.-2 + ^1^4,
()
(x
Circle
(a)
(6)
40.
(b)l.
5,
directly
from the
non-analytic?
definition.
...
directly
the
from the
. -e
definition.
l,-,-,--..-.
6-
'''''''* ''^
function.
Be1.nie ../...
(.)
Prove that
f(z)
Prove that
/(.)
= -1-
46.
47.
48.
z\^\
is
An.. (a).,^-.^-2,
49.
+ c.
,,iy,j, j
51.
Find
f{z) "^""^^
whose ""^^^
real
'^
'
nart
P^^*
= 4.-3
/'(.)
iT^^l^f^t; ^^
'^^^^^'^
^^^
.-./
e
'(xcos^
;=
's
is
such that
/(.)
- e^^^^.S^I
50.
'"
i,
LTl^,".^':
oby
and
/(I
+ , = _3.
is
^..
.,
ysiny)
the function.
(6)
/(O)
.-2,.
,(.)
^.^
_ 3. + 3 - U
52.
()
(6)
(22
+ 3) dz:
T'
Evaluate
(J+^^TiJ
^^
the circle
(6)
56.
Evaluate
Ans.
I.
(a)
() -2..-
''''" 1
aL'' 17
Ans.
1 T^f,"'^'*
+ 19i
all cases
55.
+i
where
d.
-1=3..
*o 1
2^-
+i
is:
the circle
(,)
^^1+
(h\
(^
i^Jja
(a)
1)^
+i
An.()
Ans.
{a) -8ffi/3
8,7,
'^
to 3
|.|=V2:
v^.
I
(f 2!Zd-
and then
"'^" '
''
'-''
-1 -
,"
-i +
... -2s-i
\
(6)
^^'^ ^'-^
^o
'
(c)
+ ^'-'~-,
2;rt/3
-- -losing
57.
[Hint:
Use the
definition of derivative
2^i^
"=>
Ans.
59.
(a) all
^
2,^
+i
(k\
'
(b) \z
f(
.-I
<
i?
<
2.
<
(c)
2^ (-i)n
(^2
+ 22 + 2)^
= ~li
(c)z
^"
"
-f.
60.
series converge?
niz-i)"
(n
5
-
'S
1)
()
absolutely convergent,
(6)
|.|
i.
+ i)"
^^
(
such that
1.
1.
'
Locate in the
61.
[CHAP. 17
gr^Q
g~^o
Ans.
plane
finite z
the singularities,
all
y
(h\
(a) z
(c)
z^
if
cos-,
(d)
= 1,
simple pole;
(6)
(c)
Simple poles .
= -2,
sin
/x
e)
,,v
name them:
C^)
= 0.
(e)
= ^/3,
(/)
double pole
= -1 i
z - irlZ)
,
W
'
cosz
/,!>
4- 4)2
essential singularity
removable singularity
62
^^^;
(a)
=^
,.
Ans.
(a)
and the
(c)
(z-ttY
16(z-l)
+
^
(,
Arm.
= 2;
=
2
0;
z
(a)
(b)
7/4,
-^
5!
4! z^
8/25,
65.
Evaluate
66.
If
- 9(il_l)
^
256
64
67.
If
less
^ (,+'1)^+8)
Use
<
double pole,
...,
|z-l|
7
'"
__^^
'
(x^
a"
'
""^'"^^
^'
^ ''^^^
i,
-,3Trt/2
Ans.
37r/2.
"^"''^'^
e'
f f(z) dz
0,
where
..
73.
r
J-
dx
(x^
"
+ 4f
^+
*'
J.
3a=
dx
""^
= JL
75.
32
= I
3
76.
= -5^a-', a>0
77.
'^J
^^
(x^
+ aT
8V2
r
r
Jo
+ 4)
de
cos e
277-
v/3
4^7
+ cos ef
4 cos 9
^217
r"
l)''(x^
de
(2
sin^
Vf
"
3t
d9
(1
is
at least two
DEFINITE INTEGRALS
_JL_
Ans. -S.i
show that
2V2
a''
^2vr
<
{z'
74.
z#0
poles:
r" x^dx
^^J x* + l
72
all
essential singularity,
of
z^tt
degree of P(z)
where P(z) and Q(z) are polynomials such that the
P(z)/Q(z),
EVALUATION
all
+ iy
(c) z = 2; it^ e"
z = -t;
(d)z = f,0,
f{z)
simple pole,
..
z*
= -2; 1/4
z = -5;-8/25
is
=1
RESIDUE THEOREM
64.
3)
6!
63.
i)f(,
(g-gf _
_^
4!
3)2^
4(z-ir
z^O
L-+ -i^ -
+
T i
2j
,,^
("^
RESIDUES
.2
z-r
^F
--J-zr;
/fiN
(0)
z^e-^'';
(6)
ef
2V2
f(z).
"^
''''''''
371
de
(ft
80
Jo
r"
81
a;''
6 cos 9)3
+4
"^
c^2^
ft>|6|
(^itfpp}?.
- -^-
83.
_lHl^_j^
^,-.
^0
82.
f"
Jo
l^}I}jrEax
(x^
i)^'*^
:r(2e-3)
__
= ^
-dx
'^'^''^
fsln^dx
=
85.
Jo
%
8
CO
J,
cosh
2 cosh (W2T
at (-iE 0)
(R
0^
^P
Consider
t^^"*^'
'
(I)
^i,_ ^
W;,
-^^ *='''^
"^
MISCELLANEOUS PROBLEMS
"^'''
^'
Riemanrrquatfonfire"
'"^""^'
^^'^'''^
" ^"'^
d(u, v)
then at
90.
all
points where
f'(z)
91-
^-
-^ +
If /(^)
is
dx'
w = HiA
- p
U-|
\z
a\ - R,
'"^"'"
a.
+ h)
^^
J_r
<^'
^^^^ ^^-
f^a
h)
= J^
-^ 0,
'' '
"
'"
''
^nd
/(.)d.
^r
_Mdz
^'
anA
[Hint: Write
be
R"
'^^'*
iAec>ret that
f(a
C bemg any
<^
d,f
I/'-' (a)!
prove Lawrewi's
"^ ^""^
''''^'^'
plane.
^"^^
C.
^^.
^^^^
,^
^^(^^
""' ^^^^"^
^^1^
93.
[CHAP. 17
372
Ans.
Let
:j
r" e-"F(t)
dt
/(s)
show that
pu)
J_
where
If
(see
Ans.
Chapter
(a)
is
12).
cost,
(6)
is
4+
-1
z{l
1/z)
+
^
(a)
converges for
^^''^'^
1<
l^l
-^^
+ z/2
'J
sum
_!
of degree
a given rational function with numerator
<
2)
(f e''/(z)dz
27ri t/^
/(s) is
(,
2^4
z+
Use
[Note that
_
~
-;*- -^
/(.)
(&)
of residues of
32
+ 28 + 5'
^"^
all
^ ^
e" f{z)
8(s
the poles of
/(s),
less
we can
^ -^
^
at its poles
- If
'
^^^
(?TT?
^^
'^
INDEX
Abel, integral test of, 284
Argand diagram, 6
summability, 259
theorems of, 230, 249
Absolute convergence,
Argument,
240
theorems on, 227, 239
Absolute maximum or minimum,
(see also
21
Absolute value, 3
of complex numbers, 6
Acceleration, 63, 140
centripetal, 156
in cylindrical and
Limit points)
Addition, 1
associative law of,
2, 7
Binomial coefBcients, 18
series, 231, 232
theorem, 18
Bolzano- Weierstrass theorem
12
(see
Weierstrass-Bolzano theorem)
mean value theorem, 82
Boundary conditions, 300
Boundary point, 102
Boundary-value problems,
and Fourier integrals,
328
and Fourier series, 300,
313, 314
in heat conduction,
313, 314, 328
in vibration of
strings, 319
separation of variables
Bonnet
65, 66
scale, 15, 19
system, 1
'
method for
Bounded functions,
20, 21
Arc
Bernoulli's inequality, 14
Bessel functions, 232,
250, 257 297
Bessel's differential
equation, 232 250
Binary
of elliptic integrals,
331
"tinuatio, of gamma
function, 286
Anal'^ic functions, 345
Analytic
Analytic part, of a Laurent
series, 348
Anti-derivatives, 82
110
relation,
"^
z,
to gamma functions,
287, 290
Bilinear transformation,
371 (see also
Fractional linear transformation)
eountability of, 12
Amplitud"e,r'"*''""^
y and
Bessel's inequality,
310, 320
X,
12
Algebraic numbers,
Associative law, 2, 7
for vectors, 136, 143
Asymptotic series or expansions,
233 234
'
252, 253, 259
'
for gamma function,
286, 292
Axiomatic foundations,
of complex numbers, 6
of real numbers, 3
of vector analysis, 138
Axis, real, 2
Base, of logarithms, 3
Bernoulli numbers, 258
Arithmetic mean, 9
5,
11
12
Branches of a function, 21
Branch line, 367
Branch point, 28, 348, 367
length, 94
Area, 80, 93
expressed as a line integral
204
of an ellipse, 179
of a parallelogram,
137, 148
'
373
solving.
INDEX
374
Cardinal number of the continuum, 4
Cardioid, 99
Catenary, 98
Cauchy principal value, 263, 272
Cauchy-Riemann equations, 346, 351-353
derivation of, 352
in polar form, 371
Cauchy's,
198
Constraints, 164
55, 56
recurring, 52
Continuity, 20-40, 103, 104, 111, 112, 345,
350, 351
and differentiability, 57, 63, 64, 105, 113
an interval, 25
a region, 104
350, 351
102
Closure law or property, 1
Cluster point, 5, 102 (see also Limit points)
Collection, 1 {see also Sets)
Commutative law, 2
Convergence,
test,
set,
310
equality of, 6
modulus of, 6
operations with,
6, 12,
13
6, 7,
domain
of,
228
interval of, 61
of continued fractions, 52, 53
of Fourier integrals
(see Fourier's integral theorem)
13
real and imaginary parts of, 6
roots of, 7, 13
Complex plane, 6
also Functions of
Conditional convergence,
of integrals, 262, 265, 270, 271
of series, 226
Conductivity, thermal, 314
Conformal mapping or transformation, 366
Transformations)
Conjugate, complex, 6
(see also
field,
Conservative
Constants, 4
in
region, 102
conjugate
102
in
argument
set,
Comparison
Coordinates,
curvilinear, 109 (see also Curvilinear
coordinates)
cylindrical, 142, 153, 154
hyperbolic, 185
polar, 6
INDEX
375
Coordinates (cont.)
rectangular, 6, 101, 141
spherical, 143, 153, 154
Correspondence, 2, 10, 20,
41, 101, 141
one to one, 2, 10
Dense, everywhere, 2
Countability, 4, 10, 11
of algebraic numbers,
12
of rational numbers,
105, 113
10, 11
Countable
set, 4, 10,
11
measure of a, 81, 87
Critical points, 63
Cross products, 137, 146-148
proof of distributive law
for 147
Curl, 140, 151, 152
in
curvilinear coordinates,
142
Curvature, radius of, 156,
159
Curve, coordinate, 141
simple closed, 102, 197,
204
space, 139
Curvilinear coordinates,
109
curl, divergence,
gradient and Laplacian in
m, i4^
i'
142
<* >
Jacobians and, 141, 142
multiple integrals in,
181, 182, 217, 218
orthogonal, 142
special, 142, 143
transformations and, 123,
124 141
vectors and, 141, 142
Laplacian
in, 142,
multiple integrals
parabolic, 158
volume element
in, 142,
Determinant,
for cross product, 137
for curl, 140
for scalar triple
product, 137
Jacobian (see Jacobian)
Dextral system, 136
Difference equations, 56,
285
continuous, 57
sectional or piecewise,
57
Differential equation,
Bessel's, 232, 250
Gauss', 232
solution of, by Laplace
189
in, 142,
163
Decimal representation
of real numbers,
Decimals, recurring, 1
Decreasing functions,
21, 26
monotonic, 21
approximations by use
4, 15
Definite integrals, 80,
81, 86, 87 (see also
Integrals)
in, 83,
89-92
evaluating, 85, 92
properties of, 81, 87, 88
theorem for existence of,
81
with variable limits,
83, 163, 170, 266
Degree, of a polynomial
equation, 5
o homogeneous
functions, 106
Del (V). 140
66
115
105
105
of, 65,
strictly, 21, 26
Decreasing sequences,
monotonic and strictly 42
^'
Dedekmd cuts,
change of variable
154
153
154
in,
set)
93
INDEX
376
Divergence, 140, 151
in curvilinear coordinates, 142
in cylindrical coordinates, 154
also
of improper integrals, 260-265 (see
Improper integrals)
of infinite series (see Infinite series)
Divergence theorem, 199, 200, 210-213
proof of, 210, 211
Divergent integrals, 260-265
sequences, 41 (see also Sequences)
series, 43 (see also Series)
Division, 1
by zero, 8
of complex numbers, 7, 12
Domain, of a function, 20, 101
of convergence, 228
e,
Elements, of a set, 1
Ellipse, 99
area of, 179
length of arc of, 335
also
Elliptic functions, 332, 338-340 (see
Elliptic integrals)
Empty
set, 1
Equations,
difference, 56, 285
differential (see Differential equation)
Equations
(cont.)
polynomial, 5, 21
Equipotential surfaces, 163
Errors, applications to, 164, 174
series,
in computing sums of alternating
226, 238, 239
mean square, 310
Essential singularity, 348, 358
Euler's, constant, 251, 286, 295
formulas or identities, 7, 251
theorem on homogeneous functions, 106
Even
Everywhere dense
Evolute, 176
Exact
Expansion of functions,
in Fourier series (see Fourier series)
in power
Expansions
series,
231
(see Series)
conservative, 198
scalar, 138
vector, 138
Fluid mechanics, 353
Fourier coefficients, 298
expansion (see Fourier series)
Fourier integrals, 321-330 (see also Fourier
transforms)
convergence of (see Fourier's integral
theorem)
solution of boundary-value problems by, 328
Fourier series, 298-320
complex notation for, 300
convergence
of, 299,
differentiation
311-313
and integration
of, 300,
311
INDEX
377
Frullani's integral, 282
Functions of a complex
variable (cont
Jacobians and, 371
Laplace transforms and, 372
Functional determinant
107, 120 (see also
Jacobians)
Functional notation, 20,
101
Functions, 20-40, 101, 107,
345
algebraic, 22
Bessel, 232, 250, 257, 297
beta (see Beta functions)
bounded, 20, 21
branches of, 21
composite (see Composite
contmuity of
multiple-valued, 345
poles of, 347, 348
real part of, 345, 352,
353
residue theorem for (see
Residue theorem)
series of, 347, 357-360
single-valued, 345
singular points of, 347
functions)
(see Continuity)
decreasing, 21, 26
definition of, 20, 101
Fundamental theorem,
derivatives of (see
Derivatives)
differential of (see
Differentials)
of algebra, 21
of integral calculus,
82, 88, 89
Gamma
elementary transcendental,
22 23
gamma
(see
Gamma
107
285, 287
formulas and asymptotic
series
for, 286, 292
table and graph of, 285
Stirling's
functions)
harmonic, 346
hyperbolic, 22, 23
of (see
Maxima and
minima)
monotonic, 21
multiple-valued (see
Multiple-valued
function)
normalized, 301
odd, 299, 306-309
of a complex variable
(see
complex variable)
Functions of a
of a function (see
Composite function)
of several variables,
101. 110 in
orthogonal, 301, 314,
315
orthonormal, 301
periodic, 298
polynomial, 21
sequences and series
of, 227 228 9^9
^^'
'
242, 243
single-valued, 20,
101, 345
staircase or step, 28
transcendental, 22, 23
types of, 21, 22
value of, 20
vector (see Vector
functions)
Functions of a complex
variable, oto
345-372
ai^
analytic, 345
,
Cauchy-Riemann equations
and
(see Cauchy-Riemann
equations)
imaginary
Gauss',
differential equation,
232
r function, 286
test, 227, 241
Generalized theorem of the
mean, 61, 69
treometric integral, 261
mean, 9
series, 51, 224
G.l.b (see Greatest
lower bound)
'
225
on.
INDEX
378
series, 232,
Hypergeometric function or
Hypersphere, 101
257
242, 243
of functions, 227, 228, 232,
partial sums of, 43, 224
quotient test for, 225
241
Raabe's test for, 226, 227,
241
240,
226,
for,
test
ratio
255
rearrangement of terms in, 227,
HypersTirface, 101
Hypocycloid, 207
special, 224
uniform convergence
number 6
Imaginary part, of a complex
complex variable, 345,
of functions of a
M test for,
Weierstrass
352, 353
unit, 6
Infinity, 24, 42
Integers, positive
Integrable, 81
and strictly, 42
Increaslng^st'quLces, monotonic
Indefinite integrals, 82 (^
^f"f,^JJ.
205-207, 215
198,
Independence of the path, 197,
101
Independent variable, 20,
71-74
Indeterminate forms, 33, 62,
L'Hospital's rules)
L^Hospital's rules for (see
72-74
Taylor's theorem and, 62,
14
Induction, mathematical, 7,
10
9,
2,
Inequalities,
Inequality,
Bernoulli's, 14
Bessel's, 310, 320
countably, 4
258
224-25Maeea^ao
\
c
Ser.es)
convergence
theorem of,
Integral calculus, fundamental
82, 88, 89
329
Integral equations, 324, 328,
347 353-356
Cauchy,
of
formulas
Integral
321-344, 346,
260-297,
180-223,
Integrals, 80-100,
also
(see
360-368
347, 349, 353-356,
Integration)
also Definite integrals)
definite, 80, 81 (see
double, 180
integrals)
elliptic (see Elliptic
362-366
evaluation of, 267, 275-278, 349,
Fresnel, 294
Frullani's, 282
(see
improper
Improper integrals)
indefinite, 82
for, 81, 82
contour, 349
of, 2^6
conditional convergence
tests for, 225-227
Infinite,
and negative,
188-191
SanVJi^'fiions of, 83, 89-92, 181, 182,
266, 274, 275
265,
of,
convergence
uniform
Cauchy's, 371
Schwarz's, 9, 16, 94
inferior)
Inferior limit (see Limit
Wallis', 316
Infinitesimal, 65, 79
Imaginary
of, 227,
Uniform convergence)
range
of,
special
80
methods
89-92
171
under integral sign, 163, 164,
Intercepts, 110
Interior point, 102
26
Intermediate value theorem,
Intersection of sets, 11
INDEX
379
Intervals,
closed, 4
continuity
Laplacian operator
in,
(cont.)
in curvilinear
coordinates, 142
25
m
m
infinite, 4
nested, 43, 50
of convergence, 61
open, 4
unbounded, 4
Invariance relations,
159, 160
'""tinSrr
"^^'tVsfZtr
332
''' '-'- '-'-
'''' '''
'-'-
'-^ ^^-
'
Wlaee
elliptic integrals,
331
Leibnitz's formula for
nth derivative of
hyperbolic, 23
trig-onometric, 22
Inverse, of addition
definition
103
141
Jacobi's elliptic functions,
332
addrtion formulas for,
341, 342, 344
Jacobi's forms for
332
operator)
"^'e^qttirr*^^'
"'
''-' "''"
"^^'^-'^
128, 163
102
Weierstrass-Bolzano' theorem
on
(see Weierstrass-Bolzano
theorem)
V45,'35r3^l'
definition of, 23
of a complex variable,
345, 350, 351
proofs of theorems on,
31-33
right and left hand,
23
special, 24
definition of, 41
Lemniscate, 99
Length, of a vector, 134
Less than, 2
Level curves and surfaces,
theorems on, 24
Limits of integration,
80
Limits of sequences,
of,
elliptic integrals,
331,
"''
product, 79
and multiplication
Irrational algebraic
functions 22
Irrationality of
V2, proof of, 8
Irrational numbers,
1, 2,' 8 9
approximations to, 8
Lagrange
of functions, 21
of sequences, 42,
43, 49
Left hand continuity,
25
derivatives, 57, 64, 65
limits, 23
Legendre's forms for
Inverse functions, 21
continuity of, 26
limits,
cylindrical coordinates,
142,' 154
spherical coordinates, 143
227
of functions, 227
theorems on, 41, 42, 45.47
Limits of vector
functions, 139
Limit superior, 43, 49
Linear dependence of
vectors, 160
Ivinear transformations,
131
fractional (see Fractional
linear
transformation)
of,
196
independence of path
205-207,
variable, 346
vector notation for,
196
Line, normal (see
Normal
line)
INDEX
380
Lower bound,
5, 11,
of functions, 21
of sequences, 42
Lower
Negative integers, 1
numbers, 1, 2
Neighborhoods, 5, 102
Nested intervals, 43, 50
Newton's method, 79
upper bound)
Mappings, 108
_.
..
(see also
Transformations)
conformal, 366
Mathematical induction,
Maxima and minima, 21,
7,
14
171-174
j?
icyi
,
,
for, 164,
Lagrange's multiplier method
172-174
variable, 62, 63, 74, 75
of functions of one
variables, 164,
of functions of several
171-174
relative, 21
(see
Mean, arithmetic, 9
Normal
line,
Number, cardinal,
Numbers, 1-19
numbers)
algebraic (see Algebraic
Bernoulli, 258
complex
(see
Complex numbers)
natural, 1
negative, 1, 2
operations with,
positive, 1, 2
2, 6-8, 12,
13
geometric, 9
error, 310
69
Mean, theorem or law of the, 61,
proof of, 68
Mean value theorems,
Measure
functions, 301
Normalized vectors and
numbers)
irrational (see Irrational
171, 1'^
Taylor's theorem and, 74, 75,
Mean square
vector, 135
absolute, 21
Maximum
i te
of acceleration, 156
Normal component
,
j
,
function (see Multiple-valued
M^-vaW
12
numbers)
transcendental,
5,
Numerator, 1
Numerical methods
12
.
.-
85, 92, 93
for evaluating definite integrals,
zero, 81, 87
Mechanics, 140
fluid, 353
Members, of a
Minimum
(see
Odd functions,
Open interval,
.
Modulus, complementary,
6
of a complex number,
343
331
of elliptic integrals,
210
Moebius
strip,
Moment
of inertia, 93
Monotonic functions, 21
47-49
Monotonic sequences, 42,
42
fundamental theorem on,
180-194
Multiple integrals,
improper, 267
218
coordmates, 181, 182, ill,
in curvilinear
189
coordinates,
in cylindrical
189
spherical coordinates,
in
299, 306-309
set, 1
Operations,
IS
with complex numbers, 6, 12,
231
230,
series,
power
with
with real numbers, 2, 7, 8
6
Ordered pairs of real numbers,
138
numbers,
real
triplets of
Order, exponential, 283
of derivatives, 60
of poles, 347, 348
Orientable surface, 210
101
Origin, of a coordinate system,
(see
Orthogonal curvilinear coordinates
coordinates)
Curvilinear
Orthogonal families, 352, 353
functions, 301, 314, 315
7,
12
aUo Approximations)
(see
Parallelogram, area
law, 18, 134, 144
of, 137,
148
INDEX
381
Parametric equations, of
of normal line, 161
of space curve, 139
line,
165
Power
of, 230
Prime, relatively, 8
Principal branch,
of a function, 21
of a logarithm, 351
Principal normal, to a
space curve, 155, 159
Principal part, 59, 105
of a Laurent series,
348
Parseval's identity,
for Fourier integrals,
322, 323, 325
for Fourier series,
300, 309, 310
Partial derivatives, 101-133
applications of, 161-179
definition of, 104
of,
105
Partial fractions,
84, 91 95
Partial sums of infinite
;eries, 43, 224
Pendulum, period of a
simple, 340 341
Period, of a function,
298
of a simple pendulum,
340, 341
of elliptic functions,
339, 340
Piecewise continuous, 26
differentiable,
57
to
Point
a curve
Projection, of a vector,
196
Normal plane)
' '"'"' ^^'^ ^^"^^"* P'-)
(see
boundary, 102
branch, 28, 348, 367
cluster, 5, 102 [see
also
Limit i'""ii.b;
points)
63
interior, 102
limit (see Limit
points)
neighborhood ot, 5, 102
of accumulation,
5 (see also Limit
point,)
singular (see Singular
^
"^
points)
Point set,
one dimensional, 4
normal, 199
Positive integers, 1
1,
248-250
229
special, 231,
232
theorems on, 230
series,
225
of,
I3
Quotient
test,
229, 232
ot curvature,
156, 159
of gyration,
189, 190
of torsion, 159
-Kational numbers,
1, 8,
1 (see also
Numbers)
Power
Quotient, 1
critical,
numbers,
Principal value,
of functions, 21,
22
of integrals (see
Cauchy principal value)
of inverse hyperbolic
functions, 23
of inverse trigonometric
functions, 22
of logarithms, 351
Product, 1
box, 137
cross or vector (see
Cross products)
dot or scalar (see Dot
products)
infinite (see Infinite
product)
Mth derivative of, 79
triple (see Triple
products)
Wallis', 316
Plane, complex, 6
Plane, equation of, I49
normal
series (cont.)
uniform convergence
Real part,
of a complex number,
6
* '
35? 353
"^ ^
'"'"''''''
^""^'^'' 345'
Rectangular component
vectors, 136
Rectangular coordinates,
6, 101,
141
INDEX
382
Seauences
harmonic, 225
Laurent's, 348, 359, 360, 371
Maclaurin, 61, 231
357-dbU
functions of a complex variable,
of
P-,
power
(see
sum
Sets, 1
limits, 23
<.
coordinate system,
Right handed rectangular
bounded, 5
closed, 5, 11, 12
connected, 102
Countable set)
countable or denumerable (see
135, 136
Rolle's theorem, 61
proof of, 68
elements
Roots,
13
of complex numbers, 7,
of real numbers, 3, 10
Roots of equations, 21
computations of, 36
79
Newton's method for finding,
of, 1
everywhere dense, 2
intersection of, 11
orthonormal, 301
point, 4, 101
union of, 11
204
Simple closed curves, 102, 197,
347
poles,
Simple
,. .q, ^m
197, ^04
Simply connected region, 102,
Simpson's rule, 85, 92, 93
function, 20, lOL 3^5
Single-valued
124, 260, dil,
Singular points or singularities,
357-360
34S
defined from Laurent series,
essential, 348, 358
.
Schwarz's inequality,
for integrals, 94
16
for real numbers, 9,
Section (see Dedekind
Sectional continuity, 26
differentiability, 57
integration and, 81
boundary-value
Separation of variables, in
cut)
problems, 313
Sequence, Fibonacci, 53, 55
Sequences, 41-56, 227
47-49
bounded, monotonic, 42,
41, 227
divergent,
and
convergent
increasing, 42
225
partial
infinite,
227
and
of,
series)
Region, 102
closed, 102
connected, 197
multiply-connected, 102
of convergence, 109
open, 102, 110
simply-connected, 102, 197, 204
Regular summability, 233, 258
Relativity, theory of, 160
Removable discontinuity, 34, 104
finite
terms of, 41
uniform convergence
decimal, 1
decreasing, 42
definition of, 41
(cont.)
Limits of sequences)
limits of, 41, 227 (see also
227
functions,
of
isolated, 347
removable, 348, 358
Sink, 219
Smooth function
Laplacian
in,
104
153
15^5,
in, 143,
143, 158
multiple integrals
volume element
in,
189
154
28
Staircase or step function,
Stirling's asymptotic
formula and
series, 286,
29^
INDEX
383
Stokes' theorem, 200,
213-217
proof of, 213, 214
Stream function, 353
Subset, 1
Transformations
Subtraction, 1
of complex numbers,
12
of vectors, 134
Sum,
Transforms
(see
Trigonometric functions, 22
derivatives ot, 60
integrals of, 83, 84
inverse, 22
Triple integrals,
181, 186-188
transformation
vector, 137
Unbounded interval, 4
Uniform continuity, 26, 104
Uniform convergence, 227,
equipotential, 163
level, 128,
normal
(cont.)
163
line to (see
Normal
orientable, 210
228, 243-245
of mtegrals, 265,
266, 274, 275
of power series, 230
of sequences, 227
of series, 227, 228
line)
Tangential component of
acceleration, 156
langent Ime, to a coordinate
curve 141
to a curve, 58, 162,
167, 168
Tangent plane,
161, 165-167
in curvilinear coordinates,
166, 167
Tangent vector, 139, 159
Taylor series in one variable,
61, 231 (see also
uniqueness
of,
125
approximations using, 70
for functions of one
variable, 61
for functions of several
variables, 109, 124 125
in approximate
integration, 85, 93
indeterminate forms and,
62, 72-74
proof of, 70, 125, 358
Slci''I!"''1'^''
irace, on a plane. 'f
111
Transcendental functions,
22. 23
numbers, 5, 12
Transformations, 108, 123, 124
conformal, 366
Jacobians of, 108, 142
^ t-t)
infinite dimensional,
301
rectangular, 135, 136
of functions, 20, 21
of sequences, 42
257
247
^^^-^*--
Upper bound,
variable, 347
UnToHftts, u"^^*
""' ''-'
Upper
Variable,
change
Change
4,
20
of, in diflferentiation,
59,
of, in integration,
106
'^'""
variabS
dummy, 83
limits of integration,
83, 163, 170, 266
Vector algebra, 134,
135, 143-W5
''?5o!''m
'
182
^^ '"-i^
I42
INDEX
384
(cont.)
Vectors
f 141
141
terms of,
Jacobians interpreted in
134
length or magnitude of,
normalized, 301
null, 135
position, 136
radius, 136
resultant or sum of, 134, 144
tangent, 139, 159
unit, 135, 136, 301
Volume, 94
of, 142, 143, 153, 154
149
of parallelepiped, 137, 148,
element
Weierstrass-Bolzano theorem,
50, 102
proof of, 50
test,
Weierstrass
for integrals, 266, 274-279
for series, 228, 245, 246
Work, as a
X axis, 101
intercept, 110
y axis, 101
intercept, 110
z axis,
101
intercept, 110
Zero, 1
division by, 8
measure, 81, 87
Wallis' product, 316
Wave equation, 319
5, 11, 12,
vector, 135
"7/^/
42, 43,