Definite Integration Zella - in PDF

LOCUS
Definite
Integration
CONCEPT NO
TES
NOTES
01.
Basic Properties
02.
More Properties
03.
Integration as Limit of a Sum
Maths / Definite Integration
LOCUS
Definite Integration
As explained in the chapter titled Integration Basics, the fundamental theorem of calculus tells us that to
evaluate the area under a curve y = f ( x ) from x = a to x = b , we first evaluate the anti-derivative g ( x ) of f ( x )
g ( x ) = f ( x ) dx
and then evaluate g (b ) g ( a ) . That is, area under the curve f(x) from x = a to x = b is
b
f ( x ) dx = g (b ) g (a )
a
Readers who have even the slightest doubt regarding the discussion above are advised to refer to the chapter
on Integration Basics before reading on.
Definite integration is not all about just evaluating the anti-derivative and substituting the upper and lower limits.
Working through this chapter, you will realise that a lot of techniques exist which help us in evaluating the definite
integral without resorting to the (many times tedious) process of first determining the anti-derivative. We will
develop all these techniques one by one from scratch, starting with some extremely basic properties in Section - 1
Section - 1
(1)
BASIC PROPERTIES
Suppose that f(x) < 0 on some interval [a, b]. Then, the area under the curve y = f(x) from x = a
to x = b will be negative in sign, i.e
b
f ( x ) dx < 0
a
This is obvious once you consider how the definite integral was arrived at in the first place; as a limit of the
sum of the n rectangles ( n ) . Thus, if f(x) < 0 in some interval then the area of the rectangles in that
interval will also be negative.
This property means that for example, if f(x) has the following form
y
A1
A3
b
A4
A2
Fig - 1
b
then
f ( x ) dx will equal A A
1
+ A3 A4 and not A1 + A2 + A3 + A4 .
LOCUS
If we need to evaluate A1 + A2 + A3 + A4 (the magnitude of the bounded area), we will have to calculate
x
f ( x ) dx +
From this, it should also be obvious that

(2)
f ( x ) dx + f ( x ) dx +
y
f ( x ) dx
z
f ( x ) dx f ( x ) dx
The area under the curve y = f(x) from x = a to x = b is equal in magnitude but opposite in sign to the area
under the same curve from x = b to x = a, i.e
b
f ( x ) dx = f ( x ) dx
b
This property is obvious if you consider the Newton-Leibnitz formula. If g ( x ) is the anti-derivative of
b
x(f), then
f ( x ) dx is g (b ) g ( a ) while
f ( x ) dx is g ( a ) g (b ) .
b
(3)
The area under the curve y = f(x) from x = a to x = b can be written as the sum of the area under the curve
from x = a to x = c and from x = c to x = b, that is
b
f ( x ) dx = f ( x ) dx + f ( x ) dx
Let us consider an example of this. Let c ( a, b)

y
y = f(x)
A2
A1
a
Fig - 2
It is clear that the area under the curve from x = a to x = b, A is A1 + A2.

Note that c need not lie between a and b for this relation to hold true. Suppose that c > b.
y
y = f(x)
A1
A
b
Fig - 3
LOCUS
5
b
Observe that A = f ( x ) dx = ( A + A1 ) A1
a
= f ( x ) dx f ( x ) dx
= f ( x ) dx + f ( x ) dx
Analytically, this relation can be proved easily using the Newton Leibnitzs formula.
(4)
Let f ( x ) > g ( x ) on the interval [a, b]. Then,

b
f ( x ) > g ( x ) dx.
a
This is because the curve of f(x) lies above the curve of g(x), or equivalently, the curve of
f ( x ) g ( x ) lies above the x-axis for [a, b]
y
y = f(x)
This is an example
where f(x) > g(x) > 0.
b
f (x )dx = A + A
y = g(x)
A2
while
A1
a
g (x )dx = A
Fig - 4
Similarly, if f ( x ) < g ( x ) on the interval [a, b], then

b
f ( x ) dx < g ( x ) dx
(5)
For the interval [a, b], suppose m < f(x) < M. That is, m is a lower-bound for f(x) while M is an upper
bound.
Then,
b
m (b a ) < f ( x ) dx < M (b a )
a
LOCUS
This is obvious once we consider the figure below:

y
M
C
y = f(x)
A
X
x=a
Y
x=b
Fig - 5
b
Observe that area ( rect AXYB ) < f ( x ) dx < area ( rect DXYC )
a
(6)
Let us consider the integral of f1 ( x ) + f 2 ( x ) from x = a to x = b . To evaluate the area under
f1 ( x ) + f 2 ( x ) , we can separately evaluate the area under f1(x) and the area under f2(x) and add the two
areas (algebraically). Thus:
b
( f ( x ) + f ( x )) dx = f ( x ) dx + f ( x ) dx
1
Now consider the integral of kf(x) from x = a to x = b. To evaluate the area under kf(x), we can first
evaluate the area under f(x) and then multiply it by k, that is:
b
kf ( x ) dx = k f ( x ) dx
(7)
Consider an odd function f(x), i.e, f ( x ) = f ( x ). This means that the graph of f(x) is symmetric about
the origin.
y
+
-a
a
Fig - 6
a
From the figure, it should be obvious that

right algebraically add to 0.
f ( x ) dx = 0, because the area on the left side and that on the
LOCUS
Similarly, if f(x) was even, i.e, f ( x ) = f ( x )

y
-a
+a
Fig - 7
a
f ( x ) dx = 2 f ( x ) dx because the graph is symmetrical about the y-axis.
If you recall the discussion in the unit on functions, a function can also be even or odd about any arbitrary
point x = a. Let us suppose that f(x) is odd about x = a, i.e
f ( x ) = f ( 2a x )
The points x and 2a-x

lie equidistant from
x
x=a
x = a at either sides of it.
Fig - 8
2a
Suppose for example, that we need to calculate
f ( x ) dx. It is obvious that this will be 0, since we are

0
considering equal variation on either side of x = a, i.e. the area from x = 0 to x = a and the area from
x = a to x = 2a will add algebraically to 0.
Similarly, if f(x) is even about x = a, i.e.
f ( x ) = f ( 2a x )
y
a
Fig - 9
then we have, for example

2a
f ( x ) dx = 2 f ( x ) dx
From this discussion, you will get a general idea as to how to approach such issues regarding even/odd
functions.
LOCUS
(8)
Let us consider a function f(x) on [a, b]

y
f(b)
f(a)
a
b
Fig - 10
We want to somehow define the average value that f(x) takes on the interval [a, b]. What would be an
appropriate way to define such an average?
Let fav be the average value that we are seeking. Let it be such that it is obtained at some x = c [ a, b]
y
f(b)
fav=f(c)
fav=f(c)
f(a)
Fig - 11
We can measure f av by saying that the area under f(x) from x = a to x = b should equal the area under the
average value from x = a to x = b. This seems to be the only logical way to define the average (and this is
how it is actually defined!). Thus
b
f av (b a ) = f ( x ) dx
a
f av =
1
f ( x ) dx
b a a
This value is attained for at least one c ( a, b ) (under the constraint that f is continuous, of course).
LOCUS
Example 1
1
Find the area under the curve y = 2
on its entire domain, i.e., evaluate
x +1
Solution: The graph for y =
1
dx
+1
1
is sketched below:
x +1
2
y
1
Fig - 12
Although the graph extends to infinity on both sides, the area under the curve will still be finite, as well
now see:
+
+
1
1
dx
tan
x
=
x2 + 1
= tan 1 ( + ) tan 1 ( )
What do we make of tan 1 ( + ) ? We can take it to mean lim tan 1 ( k ) which is
. Thus, the required area is .

Similarly, tan 1 ( ) would equal
2
Example 2
Evaluate the area bounded between y = x and y = x 3 from x = 0 to x = 1.
Solution: The given curves are sketched in the region of interest below.
y
y = x3
y=x
The two curves intersect
when x = x3, i.e
at x = 1 (on the positive side).
Since x3 < x for (0, 1),
the graph of x3 lies below
that of x
x=1
Fig - 13
.
2
LOCUS
10
The required area is

1
A = ( x x 3 ) dx
0
x2 x4
=
2 2 0
1 1
= (0 0 )
2 4
=
1
4
Example 3
Find the mean value of f ( x ) = cos 2 x on 0,

2
Solution: Let f av be the required mean value.
f av 0 =
2
/2
cos
x dx
1
2
/2
(1 + cos 2 x ) dx
0
/2
1
sin 2 x
= x+
2
2 0
=
f av =
1
2
Example 4
+
Evaluate
(x
Solution:
(x
+a
)( x
+ b2 )
+ a 2 )( x 2 + b 2 )
dx
can be rewritten as
1
1
1
2 2
2
2 2
(b a ) x + a x + b
2
Therefore, the given integral becomes

+
+
1
1
1
dx
dx
(b2 a 2 ) ( x 2 + a2 ) ( x2 + b2 )
+
+
1
1
1
1 x
1 x
= 2
tan
tan
a b
b
(b a2 ) a
LOCUS
11
1

2
(b a ) a b
ab ( a + b )
Sometimes, to modify an integral, an appropriate substitution has to be used; the same way we did in the unit on
Indefinite Integration. For example, integrals containing the expression (x2 + a2) can be simplified (or modified)
using the substitution x = a tan .
For evaluating a definite integral too, we can use the appropriate substitution, provided we change the limits of
integration accordingly also. This will become clear in subsequent examples.
Example 5
x/4
If I n =
tan
x dx, prove that
(a) I n + I n 2 =
1
n 1
(b)
/4
I n + I n2 =
Solution: (a)
tan n x dx +
/4
tan
1
1
< 2In <
n +1
n 1
n2
x dx
/4
( tan
x + tan n 2 x ) dx
/4
tan
x ( tan 2 x + 1) dx
n2
/4
tan
n 2
x sec 2 x dx
The substitution tan x = t can now be used to simplify this integral. However, we must change the
limits of integration according to this substitution:
tan x = t sec 2 x dx = dt
If x = 0 t = 0
t =1
If x =
4
Thus, the modified integral (in terms of the new variable t) is:
1
I n + I n 2 = t n 2 dt
0
t n 1
=
n 1 0
1
=
n 1
LOCUS
12
(b) In the integral that we are considering, the limits of integration are 0 to
In this interval, tan x < 1. Thus,

, i.e, x 0,
4
4

x 0,
4
tan n + 2 x < tan n x < tan n 2 x
From property (4), we can therefore say that:

/4
tan n + 2 x dx <
/4
tan n x <
/4
tan n 2 x
or
I n+ 2 < I n < I n2
I n + I n+2 < 2I n < I n + I n2
... (1)
Using the result of part (a) for the first and third terms in (1), we get our desired result:
1
1
< 2I n <
n +1
n 1
Example 6
x
ln t
dt. Find the function f ( x ) +
1
+
t
1
For x > 0, let f ( x ) =
1
f and show that f (e ) +
x
1 1
f = .
e 2
Solution: Observe carefully the form of the function f(x) : It is in the form of an integral (of another function), with
the lower limit being fixed and the upper limit being the variable x. As x varies, f(x) will correspondingly
vary.
One approach that you might contemplate to solve this question is evaluate the anti-derivative g(t) of
ln t
and then evaluate g ( x ) g (1) which will become f(x). However, this will become unnecessarily
1+ t
cumbersome (Try it!). We can, instead, proceed as follows:
ln t
1
f (x) + f =
dt +
x 1 1+ t
x
I1
1/ x
ln t
1 + t dt
1
I2
Notice that the limits of integration of I1 and I2 are different. If they were the same, we could have
1
added I1 and I2 easily. So we try to make them the same: in I2, if we let t = , and t varies from 1 to
y
1
, y will vary from 1 to x. This substitution will therefore make the limits of integration of I2 the same
x
as those of I1:
LOCUS
13
t=
1
y
dt =
1
dy
y2
t =1 y =1
t=
1
y=x
x
1/ x
I2 =
ln t
1+ t
dt
ln (1/ y ) 1
dy
1 + (1/ y ) y 2
1
x
=
x
=
1
ln y
dy
y (1 + y )
I1 and I2 can now be easily added:

x
ln t
ln t
= I1 + I 2 =
+
dt
+
+
1
t
t
1
t
(
)
(l+ t ) ln t
t (1 + t )
1
x
=
x
ln t
dt
t
1
dt
We used t instead of y in I 2 .
This doesn't make a difference;
y is the variable of integration;
it can be replaced with any other
variable as long as the limits of
integration are the same.
The final expression shows how simplified I1 + I 2 has become. We let ln t = z

limits of integration become 0 to ln x.
ln x
I1 + I 2 =
z dz
0
1
2
(ln x )
2
Thus,
1 ln e
f (e ) + f =
e 2
=
1
2
1
dt = dz and the
t
LOCUS
14
Example 7
/2
Evaluate
sin x + cos x
dx
Solution: The given integral can be modified into an (easily) integrable form by expressing it in a form involving
tan x and sec x.
I=
/2
1
sin x + cos x
/2
cos 2 x 1 + tan x
/2
(1 +
sec2 x
tan x
dx
dx
dx
The substitution tan x = y2 can now be used.

sec2 x dx = 2 ydy
x=0 y =0
x =
I = 2
0
y=
y dy
(1 + y )
(1 + y ) 1 dy
4
0 (1 + y )
= 2
1
1
= 2
dy
3
4
1
+
y
1
+
y
(
)
(
)
0
1 + y 2
)
(
= 2
2
1 1
= 2
2 3
=
1
3
(1 + y )
LOCUS
15
Example 8
Determine a positive integer n 5 such that
1
e ( x 1)
dx = 16 6e
Solution: Since n will not turn to be a very large integer, one might be tempted to try out various values of n in the
given relation, starting from 1 onwards, and see which one fits. This trial-and-error approach might
quickly give a result in this particular example, but what would we do it n was possibly larger?
The generally followed approach in such examples, where the integral can be characterised by positive
integer n (called the order of the integral), is to express it in terms of a lower order integral. If we
denote the nth order integral by In, we should try to express In in terms of Ik where k < n. Such a
relation is called a recursive relation. We can then simply use this relation repeatedly on any order and
obtain the integral of the next order (instead of every time repeating the calculation of integration again)
We will now use this approach on the current example:
1
Let
I n = e x ( x 1) dx
n
To simplify In, first of all let x 1 = t dx = dt and the limits become 1 to 0.

Thus,
0
I n = et +1t n dt
1
= e et t n dt
1
We now use integration by parts to solve this integral, taking tn as the first function:
I n = e t n et
n t n1 et dt dt
1
1
0
( 1)n +1
= e
n t n 1 et dt
1
e
= ( 1)
n +1
nI n 1
We have thus established the relation between In and In1 in (1)

Now observe that I0 can easily be evaluated:
0
I 0 = e et dt
1
= e 1
...(1)
LOCUS
16
Using (1) repeatedly, we can now obtain all the higher order integrals:
I1 = ( 1)
1+1
1.I 0 = 2 e
I 2 = ( 1)
2 +1
2.I1 = 5 + 2e
I 3 = ( 1)
3+1
3.I 2 = 16 6e
n = 3 is therefore the positive integer we had set out to determine. Notice the power of the recursive
relation that we obtained in (1). Using that relation, it was just a matter of minor calculations to successively
determine I1, I2 and I3 from I0. Without (1), we would have to apply integration by parts everytime,
had we used the trial-and-error approach.
Lets look at another example of this sort.
Example 9
x n
Evaluate e x dx
0
Solution: Notice that no matter what n be,

lim ( e x x n ) = 0
x
so that we will obtain a finite area under the curve.
Let
I n = e x x n dx
0
We apply integration by parts on In:

I n = x ne x
+ n x n 1e x dx
0
= 0 + nI n 1
Thus, our recursive relation is
I n = nI n1
We use this repeatedly now:
I n = nI n1
= n ( n 1) I n 2
= n ( n 1)( n 2 ) I n 3
!
= n!I0
I 0 is simple to determine:
I 0 = e x dx
0
= e x
=1
Thus,
I n = n!
LOCUS
17
Example 10
sin ( 2n + 1) x
dx = for all n "
sin x
0
Prove that
sin ( 2n + 1) x
dx.
sin x
0
Solution: Let I n =
What is I1?
sin 3 x
dx
sin x
0
I1 =
= (3 4 sin 2 x ) dx
0
= (3 2 (1 cos 2x )) dx
0
= ( x + sin 2 x ) 0
=
Thus, n = 1 satisfies the stated property. How do we approach the general case?
Some reflection on the nature of the integral will show you that evaluating In by itself would be tedius.
What we could instead do is this:
We have already shown that I1 = . If we show that I n +1 I n = 0, our task would be accomplished,
since then I 2 , I 3 and all the higher order integrals become equal to I1, which is ; this is what we
want to prove.
sin ( 2n + 3) x sin ( 2n + 1) x
I n +1 I n =
dx
sin x
sin x
= 2
0
sin x cos ( 2n + 2 ) x
dx
sin x
1
sin ( 2n + 2 ) x}
{
0
n +1
=0
Therefore,
In = n N
LOCUS
18
Example 11
/2
Evaluate
ln ( tan x ) dx
0
Solution: Careful observation will show that the function ln ( tan x ) is odd about x = , because,
4
1
ln tan x = ln ( cot x ) = ln
= ln ( tan x )
2
tan x
Thus, as discussed in property 7, the given integral will become 0.

Example 12
2
Evaluate x 2 1 dx.
0
Solution: The function to be integrated as been sketched below in the region of interest:
y
f(x)
3 2
-1
Fig - 14
This function is discontinuous; as discussed in property-3, we can split the required interval of integration.
We will do it in such a way so that in each of the sub-intervals that we obtain, the function is continuous
and can be integrated.
2
Thus, if x 1 = f ( x ) , then
2
f ( x ) dx = f ( x ) dx +
= ( 1) dx +
0
= 1 + 0 +
f ( x ) dx +
f ( x ) dx + f ( x ) dx
3
(0 ) dx + (1) dx + (2 ) dx
1
3 2 + 2 2 3
= 3 2 + 3
In general, for any discontinuous function f(x) whose integral we need to evaluate, the approach
described above is followed. f(x) is separately integrated in subintervals where it is continuous and
the results so obtained are added.
LOCUS
19
TRY YOURSELF - I
sin 2nx
dx = 0 n "
sin
x
0
Q. 1
Prove that
Q. 2
If n is an odd positive integer, prove that

/2
In =
/4
Q. 3
Show that
(cos 2 )
3/ 2
cos d =
/3
Q. 4
Show that
3
16 2
x
2
+ 2 ln
dx =
1 + sec x
18 3 3
3
x
Q. 5
sin nx
dx =
sin x
2
For x > 0, if f ( x ) =
1
ln z
1
dz , show that f ( x ) = f
z + z +1
x
2
Q. 6
1 cos nx
dx where n is a non-negative integer, show that I n , I n +1 and I n + 2 are in A.P.
1 cos x
0
If I n =
Q. 7
Show that
( x a )(b x ) dx, b > a is
equal to .
/2
Q. 8
If In =
cos n x dx, n N ,
show that I n =
Q. 9
Prove that
n 1
I n2
n
3 + 2sin x + cos x dx = 4
0
2 x5 + x 4 2 x3 + 2 x 2 + 1 ln 6 1
2 ( x 2 + 1)( x 4 1) = 2 10
3
Q. 10 Prove that
*
You must have used substitutions in some of the questions above. Think about the validity of these
substitutions. Is a substitution always valid? Or do we need to fulfill certain requirements if a substitution
is to be valid?
LOCUS
20
Section - 2
MORE PROPERTIES
In section -1 , we dealt with some basic properties that definite integrals satisfy. This section continues with the
development of some more properties that are not so trivial, and, when properly applied, can be extremely powerful.
However, even the justifications for these properties stem from obvious physical ( graphical) interpretations of
functions.
a
(9)
f ( x ) dx = f (a x) dx
0
This property says that when integrating from 0 to a, we will get the same result whether we use the
function f ( x ) or f ( a x ) . The justification for this property will become clear from the figures below:
y
y
f(x)
f(x)
a-x
f(a-x)
x
As x progresses from 0 to a,
the variable a - x progresses from a to 0.
Thus, whether we use x or a - x,
the entire interval [0, a] is still covered.
x
a
The function f(a - x) can be obtained from the

function f(x) by first flipping f(x) along the y-axis
and then shifting it right by a units. Notice that in
the interval [0, a], f(x) and f(a - x) describe
precisely the same area.
Fig - 15
There are two ways to look at the justification of this property, as described in the figures on the left and
right respectively.
Let us see how to apply this property usefully:
Example 13
Evaluate I =
/2
sin x
dx .
sin x + cos x
Solution: Observe that evaluating the indefinite integral of the function above would be very tedious. Using
Property-9, this integral can be immediately simplified; we use the substitution x
Therefore, I becomes:
I=
/2
cos x
dx
sin x + cos x
x.
2
LOCUS
21
Now, we add the original expression for I and this newly obtained modified expression:
2I =
/2
sin x + cos x
dx
sin x + cos x
/2
dx
(Quite simple now!)
I=
Example 14
Evaluate I =
/4
ln(1 + tan x) dx
0
Solution: We again use property -9 and try to simplify this integral; use the substitution x
function to be integrated:
I=
/4
x in the
4
ln 1 + tan 4 x dx
0
/4
1 tan x
ln 1 + 1 + tan x dx
0
/4
ln 1 + tan x dx
0
/4
{ln 2 ln(1 + tan x)} dx

0
/4
ln 2 dx I
0
ln 2 I
4
2I =
I=
ln 2
4
ln 2
8
Again, we see that using property - 9 saved us from the tedious task of first evaluating the
anti-derivative.
LOCUS
22
Example 15
Evaluate I =
/2
x sin x cos x
dx
sin 4 x + cos 4 x
Solution: We first use property -9 to simplify this integral; use the substitution x
integrated:
I=
/2
x cos x sin x
2
dx
4
cos x + sin 4 x
Adding the original and the modified expressions of I, we obtain :
/ 2 sin x cos x
2I =
dx
2 0 sin 4 x + cos 4 x
I=
4
/2
tan x sec 2 x
dx
1 + tan 4 x
Now we use the substitution tan 2 x = y

2 tan x sec 2 x dx = dy
and when x = 0 y = 0
y=
2
I can now be written in terms of y as :
x=
I=
dy
8 0 1+ y2
= ( tan 1 y )
8
0
2
=
16
x in the function to be
2
LOCUS
(10)
23
We can now extend property -9 into a more general property :

b
f ( x) dx = f (a + b x) dx
a
Graphically, the justification for this property is analogous to that for property -9
y
a+b-x
a+b-x
The graph of f(a+b-x) can be obtained from the

graph of f(x) by first flipping the graph of f(x)
along the y-axis and then shifting it (a+b) units
towards the right; the areas described by f(x) and
f(a+b-x) in the interval [a, b] are precisely the same
As the variable x varies from a to b,

the variable a + b - x varies from b to a.
Thus, whether we use x or a + b - x,
the entire interval [a, b] is covered in both
the cases and the areas will be the same
Fig - 16
There is a straight forward analytical justification also. Use a + b x = t in the right side integral for that
purpose. Property -10 is one of the most widely used properties to simplify definite integrals.
Example 16
Evaluate I =
/3
1+
/6
1
dx
tan x
I=
Solution:
/3
1
dx
sin
x
/6
1+
cos x
/3
/6
cos x
dx
sin x + cos x
...(1)
We now use property -10; we substitute x + x i.e. x in the function to be

6 3
integrated. Thus, I becomes:
I=
/3
/6
sin x
dx
sin x + cos x
Adding (1) and (2), we obtain
2I =
/3
/6
sin x + cos x
dx
sin x + cos x
...(2)
LOCUS
24
/3
dx
/6
I=
12
-------------------------------------------------------------------------------------------------------------------------------------2a
(11)
f ( x )dx = { f ( x) + f (2a x)} dx

0
The justification for this property is described below:

2a
2a
f ( x ) dx = f ( x) dx + f ( x) dx .
2a
To evaluate
f ( x ) dx , we can equivalently use the variable (2 a x ) instead of x, but the limits of
integration will change from (a to 2a) to (0 to a). This is because as x varies from 0 to a, 2a x will vary
from (2a to a) covering the same interval [a, 2a]. Thus,
2a
f ( x) dx = f (2a x) dx
0
Hence, the stated assertion is valid

Example 17
/2
If f is an even function, then prove that
f (cos 2 x ) cos x dx = 2
/4
f (sin 2 x ) cos x dx
Solution: On the left side, the integration limits are 0 to while on the right side, they are
2
it would be appropriate to use Property -11

/2
f (cos 2 x ) cos x dx =
/4
0 to . Thus,
4
/4
{ f (cos 2 x) cos x + f ( cos 2 x) sin x} dx

0
[since f is an even function, f ( cos 2 x ) = f (cos 2 x ) ]
/4
f (cos 2 x ){cos x + sin x} dx
= 2
/4
f (cos 2 x ) cos x + f cos 2 2 x cos 2 x dx
f (cos 2 x ) sin x + dx
4
...(1)
LOCUS
25
Now we use property - 9 to obtain the final form that we require;
use the substitution x x in the function to be integrated in (1)
4
= 2
/4
= 2

f cos 2 x sin x + dx
4
4
4
/4
f (sin 2 x ) cos x dx
We could also have started with property -9 directly:

/2
I=
f (cos 2 x) cos x dx
...(2)
/2
f cos 2 x cos x dx
2
/2
f ( cos 2 x ) sin x dx
0
/2
f (cos 2 x ) sin x dx
...(3)
Adding (2) and (3) we obtain

2I =
/2
f (cos 2 x)(sin x + cos x ) dx
Notice now that the function being integrated on the right side above is symmetric about
substitute
x for x, we will obtain the same function again. Thus, (property 7):
2
/4
2I = 2
; i.e., if we
4
f (cos 2 x )(sin x + cos x) dx
=2 2
/4
I= 2
/4
f (cos 2 x ) sin x + dx
4
f (cos 2 x)sin x + dx
4
This is the same expression that we had obtained in (1). From here, we can proceed as described
earlier.
LOCUS
(12)
26
Sometimes, it is convenient to change the limits of integration into some other limits. For example,
suppose we have to add two definite integrals I1 and I 2 ; the limits of integration for these integrals are
different. If we could somehow change the limits of I 2 into those of I1 or vice-versa, or in fact change
the limits of both I1 and I 2 into a third (common) set of limits, the addition could be accomplished
easily.
b
Suppose that I = f ( x ) dx . We need to change the limits (a to b) to (a' to b'). As x varies from a to
a
b, we need a new variable t (in terms of x) which varies from a' to b'.
t
b'
As x varies from
a to b, t varies from
a' to b'.
Thus,
t - a'
b' - a'
x-a = b-a
a'
Fig - 17
As described in the figure above, the new variable t is given by,

b a
t = a +
( x a) .
ba
Thus,
dt =
b a
dx
ba
b
I = f ( x) dx
a
b a
ba
= f a+
(t a)
dt
b a
b a
a
The modified integral has the limits ( a to b) . A particular case of this property is modifying the
arbitrary integration limits (a to b) to (0 to 1) i.e., a = 0 and b = 1 . For this case,
b
I = f ( x ) dx
a
= (b a ) f (a + (b a )t ) dt
0
LOCUS
(13)
27
If f ( x ) is a periodic function with period T, then the area under f ( x ) for n periods would be n times the
area under f ( x ) for one period, i.e.
nT
f ( x ) dx = n f ( x ) dx
0
Now, consider the periodic function f ( x) = sin x as an example. The period of sin x is 2 .
y
0 A
2 a+2
a + 2
Suppose we intend to calculate
Fig - 18
sin x dx as depicted above. Notice that the darkly shaded area in the
interval [2 , a + 2 ] can precisely cover the area marked as A. Thus,

a + 2
sin x dx
sin x dx =
This will hold true for every periodic function, i.e.

a +T
f ( x) dx = f ( x) dx
(where T is the period of f (x))
This also implies that

a + nT
b + nT
and
a + nT
b + nT
and
nT
f ( x) dx =
f ( x) dx = n f ( x) dx
0
f ( x ) dx = f ( x ) dx
a
f ( x) dx = f ( x) dx + n f ( x) dx
LOCUS
28
Example 18
n +V
Show that
| sin x | dx = (2n + 1) cos V , where n is a positive integer and V = [0, )
f ( x ) =| sin x | is periodic with period .

Therefore, as described in property-13,
Solution:
n +V
| sin x | dx = | sin x | dx + n | sin x | dx

V
sin x 0
for x [0, )
= sin xdx + n sin xdx
= cos x 0 + n( cos x ) 0
= 1 cos V + n(2)
= (2 n + 1) cos V
Example 19
n
Evaluate
1 cos 2 x dx, where n is a positive integer.
I=
Solution:
1 cos 2 x dx
2sin 2 x dx
= 2 | sin x | dx
0
= 2n | sin x | dx
(Property-13)
= 2 2n
--------------------------------------------------------------------------------------------------------------------------------------
14.
In the unit on Integration Basics, we saw that for a function f ( x ) , the anti-derivative g ( x ) was defined
as
x
g ( x ) = f (t )dt
a
so that
g '( x) = f ( x)
(where a is a constant)
LOCUS
29
We now consider an integral of the following form:
h( x ) =
(x)
f (t )dt
(x)
That is, the limits of integration are themselves functions of x. The anti-derivative g ( x ) is a special case of
h( x ) with ( x ) = x and ( x ) = a.
Now, how do we evaluate h '( x ) ? Leibnitzs rule for differentiation tells us how to do so. Since g ( x ) is
the anti-derivative of f ( x ), we have:
h( x ) =
(x)
f (t )dt
(x)
( x)
= g (t ) ( x )
= g ( ( x) ) g ( ( x) )
h '( x) = g ' ( ( x) ) '( x) g ' ( ( x) ) '( x)
= f ( ( x) ) '( x) f ( ( x) ) '( x).

Let us see an example of this rule:
Example 20
x2
Evaluate f '( x ) if f ( x) =
(t
+ 1 dt
Solution: Let us first find out f ( x ) using straight forward integration:

x2
t3
f ( x) = + t
3 x
=
x6
x3
+ x2 x
3
3
f '( x ) = 2 x 5 x 2 + 2 x 1
Now we redo this using the Leibnitzs differentiation rule:
f '( x) =
(( x ) +1)( x ) (( x) + 1)( x)
2 2
= ( x 4 + 1)(2 x ) ( x 2 + 1)(1)
= 2 x5 x 2 + 2 x 1
Let us now do an example of this rule where straight forward integration would be much more difficult.
LOCUS
30
Example 21
Determine the equation of the tangent to the curve y = f ( x ) at x = 1, where
x3
f ( x) =
dt
1+ t5
Solution: Notice how tedious it would be to actually carry out the integration. Instead, we use the Leibnitzs
differentiation rule:
f '( x ) =
f '(1) =
=
3x 2
1+ x
15
2x
1 + x10
3
2
2
2
1
2
Also, at x = 1,
1
f ( x) = f (1) =
1
1
1+ t5
dt
=0
Thus, the tangent passes through (1, 0) and has slope
1
. The required equation is
2
1
( x 1)
2
y0 =
x 2 y 1 = 0
---------------------------------------------------------------------------------------------------------------(15).
Consider a function in two variables x and y, i.e.,

z = f ( x, y )
Let us consider the integral of z with respect to x, from a to b, i.e.,
b
I = f ( x, y )dx
a
For this integration, the variable is only x and not y. y is essentially a constant for the integration process.
Therefore, after we have evaluated the definite integral and put in the integration limits, y will still remain in
the expression of I. This means that I is a function of y.
I ( y ) = f ( x, y )dx
a
...(1)
LOCUS
31
Our 15th property says that the relation (1) can be differentiated with respect to y as follows:
I '( y ) =
d
f ( x, y ) dx
dy a
f ( x, y )
dx
y
a
=
where
f ( x, y )
stands for the partial derivative of f ( x, y ) with respect to y, that is , the derivative of
y
f ( x, y ) w.r.t. y, treating x as a constant
Let us see the justification for this property:

I ( y + h) I ( y )
h 0
h
I '( y ) = lim
= lim a
f ( x, y + h)dx f ( x, y )dx
h 0
f ( x, y + h) f ( x, y )
= lim
dx
h 0
h
a
b
f ( x, y )
dx
y
a
This property turns out to be very useful in certain cases.
Example 22
1
Evaluate I =
0
xk 1
dx
ln x
Solution: Observe that I will be a function of k. Instead of carrying out direct integration, we use property -15:
1
xk 1
dI (k )
=
dx
k ln x
dk
0
1
=
0
x k ln x
dx
ln x
= x k dx
0
1
.
k +1
LOCUS
32
Thus,
dI ( k ) =
dk
k +1
Integrating both sides, we obtain

...(1)
I ( k ) = ln( k + 1) + C
To obtain C, note from the original definition of I that I (0) = 0. Using this in (1), we obtain
0 = ln1 + C
C=0
Thus,
I (k ) = ln( k + 1)
Observe again carefully the indirect route that property-15 offered us to solve this integral.
Example 23
1
Evaluate I =
0
x x
dx
ln x
Solution: We again try to use property -15 to solve this integral. Let us treat I as a function of . Therefore,
1
I ( ) =
Notice that
x x
dx
ln x
x x
I ( ) =
dx = 0
ln x
0
Now, using property-15 we obtain:
1
1
x x
dI ( )
=
dx
ln x
d
0
x ln x
dx
ln
x
0
1
=
+1
d
dI ( ) =
+1
1
Thus,
I ( ) = ln( + 1) + C
Using I ( ) = 0 above, we obtain

I ( ) = 0 = ln( + 1) + C
C = 0 ln( + 1) = ln ( + 1)
Thus,
I ( ) = ln( + 1) ln( + 1)
+1
= ln
+1
This is the required integral !

LOCUS
33
Section - 3
INTEGRATION AS LIMIT OF A SUM
In this section, we will very briefly revisit the discussion we had in the unit Integration Basics on viewing integration
as a limit of a sum. If you do not remember that discussion, you are advised to refer to it again.
In that discussion, we saw that the area under the curve is the sum of an infinitely large number of rectangles with
infinitesimally small widths. These rectangles precisely give the area under the curve in the limit that their number
tends to infinity, i.e.,:
b
hf (a + rh)
f ( x)dx = nlim
r =1
a
h=
where
ba
n
In particular, notice that if the lower limit a is 0, then

b
b br
f
n
n
r =1 n
n
f ( x)dx = lim
...(1)
We sometimes encounter series of the form as in the right hand side of (1). The discussion above shows that such
series can be summed using definite integrals.
This will become more clear through examples.
Example 24
(a)
(b)
Find the sum of the series

1
1
1
1
+
+
+ ...... +
as n
n n +1 n + 2
3n
Find the sum of the series
1
1
n
+ 2
+ ...... + 2
as n
2
2
n +1 n + 2
n + n2
2
(c)
Evaluate
[(n + 1)(n + 2).....(n + n)]1/n

n
n
lim
Solution: (a) The given series can be written concisely as

2n
1
S =
r =0 n + r
=
1 2n
1
n r = 0 1 + ( r/ n )
Comparing this with the right hand side of (1), we see that S can be expressed as the integral of a
1
function f ( x) =
from 0 to 2, because since r varies till 2n, r/n varies till 2. Thus,
1+ x
2
1
dx
1+ x
0
S=
= ln 3
LOCUS
34
(b) Concisely put,

n
n
2
l =1 n + r
S =
=
1 n
1
n r =1 1 + ( r/ n) 2
As discussed earlier, S can be written in integral form as

1
1
dx
2
1
+
x
0
S=
= tan 1 x
=
1
0
(c) Let L represent the given limit. We have,

ln L =
1 1 2
n
ln 1 + 1 + ...... 1 +
n n n
n
1 n r
= ln 1 +
n r =1 n
Thus,
ln L = ln(1 + x)dx
0
x
dx
1+ x
0
= x ln(1 + x) 0
1
x +1 1
= ln 2
dx
+
1
x
0
1
= ln 2 1 + ln 2
= 2 ln 2 1
= ln(4/e)
This implies that

4
e
From these three examples, the usefulness of definite integrals in summing series should be quite
apparent.
One last thing about definite integration as the limit of a sum form: when we divide the area we want to evaluate
into n rectangles, we need not have those n rectangles of the same width. The widths can be arbitrary as long as
all of them tend to 0 in the limit n . For example, we could divide the interval [a, b] as follows:
L=
To divide [a, b] into n

a
ar
ar2
ar ...........b=arn
Fig. 19
intervals, we let arn=b.

ln (b/a)
Thus, n =
ln r
As r 1, n
LOCUS
35
In the scheme of division above, as r 1, n . This is also a valid division scheme since the width
of the kth interval is ar k ar k 1 which tends to 0 as r 1 or n .
In fact, sometimes a uniform division scheme might not work due to the nature of the function to be
integrated. In such cases, the division scheme described above could be used.
Example 25
b
1
Evaluate the integral dx using first principles.
x
a
Solution: As described earlier, we divide the interval [a, b] into n non-uniform sub intervals, the kth interval being
of width wk = ar k ar k 1 = ar k 1 (r 1). The sum of the areas of the n rectangles is
n
S = wk f (ar k 1 )
k =1
n
= ar k 1 (r 1)
k =1
1
ar k 1
= (r 1)
k =1
= n ( r 1)
(r 1) ln(b/a )
ln r
p ln(b/a )
ln( p + 1)
Thus,
lim S = lim S = lim S = ln(b/a )
r 1
p 0
x dx = ln a
where p = r 1;
as r 1, p 0
LOCUS
36
TRY YOURSELF - II
Q. 1
x sin x
2
dx
=
Prove that
4
2 sin 2 x
0
Q. 2
x 2 cos x
dx = (2 )
Prove that
2
+
(1
sin
)
x
0
Q. 3
Prove that ln(1 + cos x)dx = ln 2

0
Q. 4
Prove that
1 + sin sin x dx = cos 2

0
x2
Q. 5
If y =
ln(1 + x)dx,
0
show that
dy
= 2 x ln(1 + x 2 )
dx
Q. 6
Prove that
([ x] + [ x]) dx = a b
a
/2
Q. 7
Show that
sin 2 x ln(tan x)dx = 0

0
Q. 8
2
x
=
dx
Show that
1 + cos 2 x
2 2
0
1
Q. 9
1
n
1
If I n = x tan x dx, prove that (n + 1) I n + (n 1) I n 2 =
2 n
0
(n !)1/n
n
n
Q. 10 Evaluate lim
LOCUS
37
SOLVED EXAMPLES
Example 1
n/2
Evaluate
x sin x cos x
dx where a, b > 0
(a cos 2 x + b 2 sin 2 x) 2
2
sin x cos x
( a cos 2 x + b 2 sin 2 x ) 2
Solution: Consider only the function f ( x) =
Observe that f ( x ) can be easily integrated. Using the substitution sin 2 x = t , we get the integral as
(verify):
1
f ( x)dx = 2(b 2 a 2 )(a 2 cos2 x + b 2 sin 2 x)
Now since we know the integral of f (x), we use integration by parts on the original integral, which is
nothing but:
I=
/2
x f ( x )dx
Applying integration by parts, we obtain:
I = x f ( x)dx
/2
0
/2
( f ( x) dx ) dx
0
1
x
=
+
2
2 2
2
2(b a ) a cos x + b 2 sin 2 x 0
/2
1
dx
a 2 cos 2 x + b 2 sin 2 x
/ 2
1
sec 2 x
=
+
dx
2(b 2 a 2 ) 2b 2 0 a 2 + b 2 tan 2 x
To evaluate this integral, use the substitution tan x = t ;
dt
this integral reduces to 0 2 2 2 which is 2ab
a +b t
Example 2
1
Evaluate I = ln
0
1 x + 1 + x dx

1

2
2
2(b a ) 2ab 2b
.
4ab (a + b)
LOCUS
38
Solution: The expression for I contains both (1 x ) and (1 + x ); the substitution x = cos 2 could do the job.
x = cos 2
dx = 2sin 2 d
when x = 0, = /4
when x = 1, = 0.
Thus, I gets modified to
ln (
0
I = 2
1 cos 2 + 1 + cos 2 sin 2 d
/4
/4
=2
ln { 2 (sin + cos )}sin 2 d

0
/4
/4
=2
+ cos ) sin 2 d
(ln 2 ) sin 2 d + 2 ln( sin First
Second
= ln 2(cos 2 )
/4
0
Function
Function
ln(sin + cos ) cos 2

+ 2
2
/4
+
0
/4
cos sin cos 2
sin + cos
2
1 /4
= ln 2 + 2 0 + (cos sin ) 2 d
2 0
= ln 2 +
/4
(cos sin )
= ln 2 +
/4
(1 sin 2 )d
0
= ln 2 +
= ln 2 +
cos 2
+
4
2
/4
0
4 2
Example 3
1
If m, n ", prove that I m ,n = x m (1 x) n dx =

0
m!n!
(m + n + 1)!
Solution: It is clear from the final result which we need to obtain that we require a recursive relation involving
I m ,n and a lower order integral (w.r.t one of the two variables m or n).
LOCUS
39
1
I m ,n = (1 x ) n x m dx
0
IInd
Function
Ist
Function
(1 x) n x m +1
n
n 1 m +1
=
+
(1 x) x dx
m +1
m
+
1
0
0
= 0+
n
I m +1, n 1
m +1
Thus, the required recursive relation is

I m ,n =
n
I m +1,n 1
m +1
We use this relation repeatedly now till n reduces to 0:

I m ,n =
=
n
I m +1,n 1
m +1
n( n 1)
I m + 2, n 2
( m + 1)( m + 2)
$
n(n 1).............1
I m + n,0
(m + 1)(m + 2)....(m + n)
m !n !
I m + n,0
(m + n)!
I m + n ,0 is easy to evaluate. Verify that it is
Thus,
I m, n =
1
m + n +1
m !n !
(m + n + 1)!
Example 4
1
2
1
Evaluate I = cot (1 + x x)dx
0
Solution: Before reading the solution below, you are advised to try simplifying I directly using the properties and
techniques developed in this chapter. Youll realise the difficulty of the task.
LOCUS
40
In this example, instead of direct application of any property, a non-trivial manipulation is first required
to simplify I before applying any properties. That manipulation is now described:
1
I = cot 1 (1 + x 2 x)dx
0
1
= tan 1
dx
1 + x( x 1)
0
1
x ( x 1)
= tan 1
dx
1 + x( x 1)
0
1
We wrote the numerator '1'
as 'x ( x 1) '
= tan 1 x tan 1 ( x 1) dx
0
1 A B
1
1
tan
= tan A tan B
1 + AB
= tan 1 x dx tan 1 ( x 1)dx
...(1)
Observe how simple the integral I has now become. A further simplification is possible by the application
of property - 9 on the second integral in (1):
1
1
1
tan ( x 1) = tan ((1 x) 1) dx
1
= tan 1 x dx
0
I = 2 tan 1 x dx
Thus,
We can now proceed to evaluate I using integration by parts:

1
x
1 1
I = 2 x tan x
dx
0
1 + x 2
0
1 2 dt
= 2
4 2 1 t
= (ln t )
2
1
ln 2
2
We used the substitution
2
1 + x = t
LOCUS
41
Example 5
x sin 2 x sin cos x

2
dx
Evaluate I =
2x
0
Solution: By property - 9, we use the substitution x x in the function to be integrated:
( x ) sin(2 2 x ) sin cos ( x )

2
dx
I =
2( x )
0
( x) ( sin 2 x) sin cos x

2
dx
=
2x
0
cos x
( x )sin 2 x sin
2
dx
=
2x
0
Adding the original and the modified forms of I, we obtain:
cos x
(2 x ) sin 2 x sin
2
dx
2I =
2x
0
= sin 2 x sin cos x dx

2
0
We use the substitution

to
:
2
2
cos x = t in this simplified integral. The limits change from (0 to ) to

2
2 I = 2 sin x cos x sin cos x dx

2
0
=
8
2
8
= 2
t sin t dt
/2
/2
/2
t sin t dt
/2
8
/2
+
t
cos
t
cos t dt
2
2

/2
8
= 2 {0 + 2}
16
= 2
8
I= 2
- /2
LOCUS
42
Example 6
Evaluate I =
2 x(1 + sin x)
dx
1 + cos 2 x
Solution: The limits of integration for I are symmetric about x = 0; this suggests that we should look for even/odd
portions present, if any, in the function to be integrated:
odd
even
x sin x
x
dx + 2
dx
2
2
+
+
1
cos
1
cos
x
x
I = 2
x sin x
dx + 0
1 + cos2 x
0
= 4
...(1)
( x) sin( x)
dx
2
1
+
cos
(
x
)
0
= 4
( x) sin x
dx
1 + cos 2 x
0
= 4
...(2)
Adding the expressions for I in (1)and (2), we obtain:
sin x
dx
2
+
1
cos
x
0
2 I = 4
Now use the substitution cos x = t; sin x dx becomes dt and the limits of integration change from (0
to ) to (1 to 1):
1
dt
1+ t 2
1
I = 2
= 2 (tan 1 t )
1
1
=2
Example 7
Prove that I =
x sin 2 n x
dx = 2 .
2n
2n
sin x + cos x
Solution: Applying property -9 on I and then adding the original and modified forms of I, we obtain:
2 I = 2
sin 2 n x
dx
sin 2 n x + cos 2 n x
LOCUS
43
The function to be integrated is symmetric about (verify this by the substitution x2 x). Therefore:
sin 2 n x
dx
2n
2n
+
sin
x
cos
x
0
2 I = 4
The function to be integrated is still symmetric, about x =

Therefore:
2 I = 8
/2
; verify this by the substitution x x.

2
sin 2 n x
dx
Now we apply property -9 on the integral above, i.e., we use x
2 I = 8
/2
cos 2 n x
dx
...(1)
x. Thus,
2
...(2)
Adding (1) and (2), we obtain
4 I = 8
/2
= 8
dx
/2
dx
0
= 4 2
I =2
Example 8
Find a lower bound and an upper bound for the integral
1
I =
0
1
1 + x6
dx
Solution: By this question, we mean that we need to find two values m and M between which I lies, i.e.,
m < I < M.
Technically, any number less than I, say (1 billion) is a valid lower bound for I and any number
greater than I, say (1 billion), would be a valid upper bound. But how useful would it be to state the
(trivial) fact that I lies between ( 1 billion) and (1 billion)? Not much.
We want tight bounds, i.e., narrow ranges in which I could lie. Thus, M m should be as small as
possible so that we have an accurate idea about the approximate value of I.
LOCUS
44
Lets obtain an upper bound first:

We have,
1 + x6 > 1
1 + x6 > 1
1 + x6
1
<1
for
x (0,1)
for
x (0,1)
for
x (0,1)
for
x (0,1)
for
x (0,1)
1
1 + x6
dx < 1 dx
0
I <1
Thus we now know that I is less than 1.

Lets obtain a good lower bound now:
Since x (0,1),
x 6 < x. Thus,
1 + x 6 < 1 + x < (1 + x ) 2
1 + x6 < 1 + x
1
1
1 + x6
1
dx
1+ x
0
dx >
I > ln 2
Thus, we now have a fair idea about the approximate value of I:

ln 2 < I < 1
Try to obtain tighter bounds for yourself.
Example 9
Evaluate I n =
/2
sin
x dx
Solution: Let f n ( x ) = sin n x dx
f n ( x ) dx = sin n x dx
= sin n 1 x sin x dx
Ist
function
IInd
function
= sin n 1 x cos x + (n 1) sin n 2 x cos 2 x dx

= sin n 1 x cos x + ( n 1) sin n 2 x(1 sin 2 x) dx
LOCUS
45
= sin n 1 x cos x + (n 1) f n 2 ( x) (n 1) f n ( x)
nf n ( x ) = sin n1 x cos x + (n 1) f n2 ( x)
sin n 1 x cos x n 1
+
f n ( x) =
f n 2 ( x)
n
n
/2
n 1
I n = f n ( x) 0 = 0 +
I n2
n
Thus, our recursive relation is
n 1
In =
I n2
n
We now use this repeatedly:

n 1
In =
I n2
n
=
( n 1)( n 3)
I n4
n( n 2)
$
(n 1)(n 3).............(2)
n(n 2)....................(3) I1
=
or
(n 1)(n 3).............(1)
I0
n( n 2)....................(2)
I1 =
Now,
(if n is odd)
(if n is even)
...(1)
/2
sin x dx = 1
0
I0 =
and
/2
dx = 2
0
Thus, I n is now obtainable from (1).

Notice that J n =
/2
cos
x dx will be the same as I n , by virtue of property 9.
Example 10
Suppose that a + b = 4 where a (0, 2) and g ( x ) is a differentiable function such that g '( x ) > 0 V x % .
a
Show that g ( x )dx + g ( x)dx increases as (b a) increases:
LOCUS
46
Solution: Let us first intuitively try to justify the stated assertion.

Since g '( x ) > 0 V x % , g ( x ) is an increasing function on !.
Now, since 0 < a < 2 and a + b = 4, a and b will lie symmetrically about the point x = 2. As a
increases from 0 to 2, b decreases from 4 to 2. Assume an arbitrary (increasing) configuration for
g ( x) :
y
g(x)
a and b lie symmetrically
about the point x = 2;
a(0, 2) and b(2, 4)
x
a
b4
Fig - 20
a
Now we interpret the area given by g ( x )dx + g ( x)dx from the figure below
0
g(x)
A1
A2
a
A3
b
Fig - 21
We have
A = g ( x)dx + g ( x)dx
= A1 + ( A1 + A2 )
= 2A1 + A2
= ( A1 + A2 + A3 ) ( A3 A1 )
We introduced A3 into
the expression
= g ( x)dx ( A3 A1 )
...(1)
Now consider what will happen as (b a) increases. The area A2 expands, while A1 and A3 will
shrink (visualise this expansion and shrinking of areas from Fig.21). But the important point is that
the shrink in A3 and A1 will not be the same. Since g ( x ) is increasing, A3 will shrink more than A1
will. Thus, as (b a ) increases, ( A3 A1 ) will decrease.
LOCUS
47
From (1), therefore, A will increase!

The maximum A is obtained when A3 A1 = 0, i.e.,
When a = 0 and b = 4:
4
Amax = g ( x)dx
0
Now we redo this example analytically. We treat a as a variable which can lie in (0, 2). Also,
b = 4 a.
Thus,
a
4a
A = g ( x)dx + g ( x) dx
= g ( x)dx +
g ( x) dx
dA
= g ( a ) g (4 a )
da
Since g ( x ) is an increasing function and a < 4 a, we have g ( a ) < g (4 a ) so that

dA
<0
da
Thus, A decreases with respect to a or equivalently, increases with respect to b a.

Example 11
Evaluate these limits:
x
(1 cos 2 x) dx
(a) L1 = lim 0
x 0
(b)
x tan x dx
x x
e dx
L2 = lim 0x
x 0
x
e dx
2
0
, and therefore, we can use the L.H. rule.
0
differentiate the integrals, we can use the Leibnitzs differentiation rule.
Solution: Observe that both these limits are of the from
(a)
L1 = lim
x 0
1 cos 2 x
x
x tan x + tan x dx
0
still of the form

0
To
LOCUS
48
2sin 2 x
x 0 x sec x + tan x + tan x
= lim
4sin x cos x
x 0 x sec 2 x + 2 tan x
= lim
4sin x cos3 x
x 0 x + 2sin x cos x
= lim
4 cos3 x
x 0 x
+ 2 cos x
sin x
= lim
(b)
4
1+ 2
4
.
3
x x x
2 e dx .e
L2 = lim
4
x
x 0
2 xe
still of the form

0
x
e dx
x
e
= lim 4 lim 0
x 0
e x x 0 x
The second limit is of
the form 0
ex
x 0 1
= 1 lim
=1
Example 12
Find a function g ( x ) continuous in (0, ) such that g ( x ) > 0 V x (0, ) and g (0) = 0, and
x
2
=
g
t
dt
(
)
g (t ) dt
x0
0
Solution: Differentiating the given relation w.r.t. x, we obtain:

x
x
4
2
g ( x) = g (t )dt g ( x) 2 g (t )dt
x 0
x 0
LOCUS
49
x
Let
g (t )dt = y. Thus,
0
g 2 ( x) =
4 y g ( x) 2 y 2
2
x
x
2 y 2 4 xg ( x ) y + x 2 g 2 ( x ) = 0
y=
4 xg ( x ) 16 x 2 g 2 ( x) 8 x 2 g 2 ( x)
4
(4 2 2 ) xg ( x)
4
= 1
x g ( x)
2
Thus,
k = 1
g (t )dt = kx g ( x)
0
Differentiating both sides, we obtain:
g ( x) = k {g ( x) + xg '( x)}
kx d g ( x )
dx
(1 k ) g ( x ) =
dg ( x) 1 k dx
=
g ( x) k x
Integrating both sides, we obtain,

1 k
ln g ( x ) =
ln x + C
k
g ( x) = C0
1 k
x k
Take C as ln C0 where
C0 is another constant
Since g (0) = 0, the power of x must be positive, i.e.,

k = 1
1
2
Thus,
g ( x) = C0 x
1
2 1
= C0 x
2 +1
LOCUS
50
Example 13
Prove that if k & + ,
sin 2kx
= 2 [cos x + cos 3 x + ........ + cos(2k 1) x ]
sin x
Hence or otherwise, prove that

/2
sin 2kx cot xdx = /2

0
Solution: Let us first directly try to prove the second part using the technique of recursion.
Let I k =
/2
sin 2kx cot x dx

0
I1 =
/2
sin 2 x cot x dx
0
/2
= 2 cos 2 x dx
0
/2
1 + cos 2 x
=2
dx
2
0
=
Also,
I k +1 I k =
.
2
/2
(sin(2k + 2) x sin 2kx ) cot x dx

0
/2
= 2 cos(2k + 1) x sin x cot x dx

0
/2
(2 cos x cos(2k + 1) x ) dx
0
/2
(cos(2k + 2) x + cos 2kx ) dx

0
/2
/2
sin(2k + 2) x
sin 2kx
2k + 2 0
2k 0
= 00
=
=0
Thus,
Ik =
for all k & +

2
LOCUS
51
Now well use the first result mentioned in the question to prove the second part.
The proof of the first result is simple:
2sin x [cos x + cos 3 x + .......... + cos(2k 1) x ]

= sin 2 x + (sin 4 x sin 2 x ) + ..... + (sin 2kx sin(2k 2) x )
= sin 2kx
Thus, the stated assertion is valid

Now,
I=
/2
sin 2kx cot x dx

0
/2
= 2 cos x [cos x + cos 3x + .......cos(2k 1) x ] dx

0
Using the
first result
/2
{1 + cos 2 x + (cos 4 x cos 2 x) + ...... + cos 2kx cos(2k 2) x}dx

0
/2
(1 + cos 2kx)dx
0
{cos 2kx integrates to 0}
Example 14
x
Suppose that g ( x ) is an even function, and f ( x) = g (t ) dt.

0
Is f ( x ) even or odd, or neither?
f ( x ) =
Solution:
g (t )dt
0
If we let t = y, the limits of integration change from (0 to x) to (0 to x). Thus,

x
f ( x ) = g ( y )( dy )
0
= g ( y )dy
0
= f ( x)
Thus, f ( x ) is odd.
( g is even)
LOCUS
52
Example 15
x
Let x > 0. Evaluate [t ]dt.

0
Solution: The function to be integrated is discontinuous at all integer points. Therefore, we integrate it piecewise:
x
[t ]dt = [t ]dt + [t ]dt + .....

0
[ x]
[t ]dt +
[ x ]1
[t ]dt
[ x]
[ x]
[ x ]1
= 0 dt + 1 dt + 2 dt + ..... + ([ x] 1)
= 0 + 1 + 2 + ....... ([ x] 1) + ( x [ x ])([ x])
=
[ x ] ([ x ] 1)
+ [ x ] ( x [ x ])
2
dt + [t ]dt
[x]
Verify that the last
term is correct

Definite Integration Zella - in PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Definite Integration Zella - in PDF

Uploaded by

Copyright:

Available Formats

LOCUS

Integration as Limit of a Sum

Maths / Definite Integration

Maths / Definite Integration

From this, it should also be obvious that

Let us consider an example of this. Let c ( a, b)

It is clear that the area under the curve from x = a to x = b, A is A1 + A2.

Maths / Definite Integration

Let f ( x ) > g ( x ) on the interval [a, b]. Then,

Similarly, if f ( x ) < g ( x ) on the interval [a, b], then

Maths / Definite Integration

This is obvious once we consider the figure below:

Let us consider the integral of f1 ( x ) + f 2 ( x ) from x = a to x = b . To evaluate the area under

From the figure, it should be obvious that

f ( x ) dx = 0, because the area on the left side and that on the

Similarly, if f(x) was even, i.e, f ( x ) = f ( x )

f ( x ) dx = 2 f ( x ) dx because the graph is symmetrical about the y-axis.

The points x and 2a-x

x = a at either sides of it.

Suppose for example, that we need to calculate

f ( x ) dx. It is obvious that this will be 0, since we are

then we have, for example

Let us consider a function f(x) on [a, b]

Maths / Definite Integration

Solution: The graph for y =

What do we make of tan 1 ( + ) ? We can take it to mean lim tan 1 ( k ) which is

. Thus, the required area is .

Maths / Definite Integration

The required area is

Find the mean value of f ( x ) = cos 2 x on 0,

Therefore, the given integral becomes

Maths / Definite Integration

x dx, prove that

tan n + 2 x < tan n x < tan n 2 x

From property (4), we can therefore say that:

For x > 0, let f ( x ) =

Maths / Definite Integration

I1 and I2 can now be easily added:

The final expression shows how simplified I1 + I 2 has become. We let ln t = z

Maths / Definite Integration

The substitution tan x = y2 can now be used.

Maths / Definite Integration

To simplify In, first of all let x 1 = t dx = dt and the limits become 1 to 0.

We have thus established the relation between In and In1 in (1)

Solution: Notice that no matter what n be,

so that we will obtain a finite area under the curve.

We apply integration by parts on In:

Maths / Definite Integration

Thus, as discussed in property 7, the given integral will become 0.

If n is an odd positive integer, prove that

( x a )(b x ) dx, b > a is

Maths / Definite Integration

The function f(a - x) can be obtained from the

Maths / Definite Integration

(Quite simple now!)

{ln 2 ln(1 + tan x)} dx

Maths / Definite Integration

Adding the original and the modified expressions of I, we obtain :

Now we use the substitution tan 2 x = y

Maths / Definite Integration

We can now extend property -9 into a more general property :

The graph of f(a+b-x) can be obtained from the

As the variable x varies from a to b,