Student Solutions Manual and Study Guide

Chapters 1 & 2 Preview

for
Numerical Analysis
9th EDITION

Richard L. Burden
Youngstown State University

J. Douglas Faires
Youngstown State University

Prepared by

Richard L. Burden
Youngstown State University

J. Douglas Faires
Youngstown State University

Australia • Brazil • Japan • Korea • Mexico • Singapore • Spain • United Kingdom • United States
© 2011 Brooks/Cole, Cengage Learning ISBN-13: 978-0-538-73563-6
ISBN-10: 0-538-73563-5
ALL RIGHTS RESERVED. No part of this work covered by the
copyright herein may be reproduced, transmitted, stored, or Brooks/Cole
used in any form or by any means graphic, electronic, or 20 Channel Center Street
mechanical, including but not limited to photocopying, Boston, MA 02210
recording, scanning, digitizing, taping, Web distribution, USA
information networks, or information storage and retrieval
systems, except as permitted under Section 107 or 108 of the Cengage Learning is a leading provider of customized
1976 United States Copyright Act, without the prior written learning solutions with office locations around the
permission of the publisher except as may be permitted by the globe, including Singapore, the United Kingdom,
license terms below. Australia, Mexico, Brazil, and Japan. Locate your local
office at: international.cengage.com/region

Cengage Learning products are represented in

For product information and technology assistance, contact us at Canada by Nelson Education, Ltd.
Cengage Learning Customer & Sales Support, 1-800-354-
9706 For your course and learning solutions,
For permission to use material from this text or product, submit all
visit academic.cengage.com
requests online at www.cengage.com/permissions Further
Purchase any of our products at your local
permissions questions can be emailed to
college store or at our preferred online store
permissionrequest@cengage.com
www.ichapters.com
Contents
Prefae v
Mathematial Preliminaries1 1
1 23
Solutions of Equations of One Variable
9
15
12
34
2 56
2814
Interpolation and12Polynomial Approximation 37
2
34 4 40
35 58
Numerial Differentiation1 and Integration
23 615
45 7048
67 14
4 89 6

Initial-Value Problems1 for Ordinary

Equationsiii...... . . . . . . . . . . . . . . . . . . . . .
8941

Differential Exerise Set 5.23 . . . . . . . . . . . . . . . . .

iv 4
5
6
7
97
3
4
8
8 09
9 . 1
0 4
5 1 5
Diret Methods for1 Solving Linear Systems 19
3 29
2
66 43
4
5 7
Iterative T hniques in Matrix Algebra 1
1
2 3
3 5
4 7
5 58
76 0
Approximation Theory
1 5
2 67
3 0
4 1
5 3
86 5
Approximating Eigenvalues 79
4

2
1
2
3 4
88
9.65 .
3
Numerial Solutions of Nonlinear Systems of Equations 19920
Exerise Set 10.5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
2
3 4
4
7
Boundary-Value Problems for Ordinary Differential Equations v
1 3
2 6
3 19
4 2
1 5 25

Numerial Solutions to Partial Differential Equations

1
2
3
3 39

Exerise Set 12.4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 242
vi Prefae
Prefae
This Student Solu Manual and S udy for Numerial Analysis, Ninth Editio , by Burden Faires
o k. attention was paid to en ure that the exerises solved in the Gui tehare requiring
into thePartiularory ed in thGuidebook. Although answers to theodd exeris disussearealsoinsightthe
system (CAS). W methodshoM apl
as our standard CAS, beausthe their pakage p llels

ntains ative exerises that have been worked out in detail for all niques and
bak f epresenbook, the re ults list in this Stu y Guide generally go well tho thosein book.
For this edition we have adddisus number exerises to the xt that involve the use of a ompu er

the algorit ms in this book. However, any of ommon omputer alg ialAnalysissystems,suh Mathemlgebratia,
ourse we have found that understood theonepts better when beyondworked through the algorithms
It has been our pratiedomaintolude strutured algorithms in ur Numerial Analysis book for all

MATLAB, and
prog ram ming lang uag e, b ystudentsts with even
he publi syst m, Sage, an be used with satisfation. In our reent teahing of mi imal a mount of prog ra mming ex pertise.

step-by- but let omputer algebra system do the tedious .

tehniqustep,disussed in the ext. The algorithms are given inomputationfrmthat an be od d in any appropriatthe

At the website for the book, supple eth avex erisdtheM apr eg ularlyawar eNum erialents,e,andPas al,sug g estionsM ATLof,toapprreasAC,welletnal ysisB,priateJa va anyb easd.

lwilltionsgebraanguagesany.beerratatoofsyupdatrethetems,FORTRAN,w
Analysis
modipakageMathematiaYouyThehttp://wwwillationswebandalsondtoofthe.qitodethatu.Forndthemathestionsontainsnumerousmightftehniquesthis.deysuallfromeditionin.beedu/additionalthethelpshangmadealgorithmsusers.formwefaires/Numerial.haveofForofthatinformationtheworksheetsexample,haverewrittenbookbeenonernin-aboutAnalysis/weforallmadetheeditionswilltheprogrammingthengtoMomputerplaebook,thisinterpretplesystemthereandprograms
appliationsrespons es ∼

We hope r Gu de you with your study of Numerial . If you any for

improv ments that an be in rporated into future of the book or would
listed below.
most grateful to ree v your omments. We an be most ea ily ontated by theeltroni mail at weth addresses

Youngstown State University Rihard L. B rden

burdenmath.ysu.edu

August 14, 2010 J. Douglas Faires

 fairesmath.ysu.edu

vii
viii Prefae
Mathematial Preliminaries
Exerise Set 1.1, page 14
2. . Find intervals that a solutionxmusttotheexistequationin(4, 5) with 0 = f (x) = x − (ln x) .
x
1. d.SOLUTION:ShowthattheIt isequationnotpossible to lgebraially hasatsolveleastfornethesolutionin the in erval x −(ln x )=0 [4, 5].

problem, we must show only that a solution exists. Let x , butthis is not required in the

Sine f (x) = x − (ln )x = x − exp(x(ln(ln x))).

an ur
x = − 32 and x = 2. maximum ly at
Intermediate Value
Theoremfis1ontinuous.11impliesontainthat [4,5]numberwithf (4) ≈ 0.3066 and f (5) ≈ −5.799, the x

SOLUTION: Let x3 − 2x2 − 4x + 3 = 0 .

f (x ) = x . The ritial points of our when
3 − 2x 2 − 4x + 3 f

whih lies in the interval ′ R lative

f (x ) = x
(e + 2) /3 , soandtheminimumonlyritialvaluespointof ours at ,
of f x = ln 2

at is, when 0=f ′

(x) = 3x2 −4 − 4 = (3x + 2)(x − 2);

[0 1] . The maximum for ; |f (x)| must onsequently be

threese.valuesSine. There are at mostintervalsthresolutions to , beause andis polynomial of f (x ) = 0 f (x )
egree
2
f (−2) = −5 and f − 3 ≈ 4.48 (0) = 3 and f (1) = −2; f (2) = −5 and
4. f a. Find ; solutions lie in the
(4) =19 , and
[−2, − 2/3] [ 0, 1], [2, 4].

SOLUTION:(x)=x+ 2Forx+ k ross s the x-axis exatly one, regardl k

SOLUTION: First notewhenthat .
−
max f (x ) f (x ) = (2 + 2x ) / 3

0 x | |
1 1 −2
≤ ≤
x < 0,we f (x ) < 2x + k

−
′ ′
5. IfandUsef(cthe)=Intermediate0ndf(c ) Value=0forT someoremc 1.11=c,andthenRolbylTheor'sTheor 1.there7,erem1.7to showexitsthatanumberthegraph pbetweenof c
max{|f (0)|, |f (ln 2)|, |f (1)|} = max{1/3 (2 ln 2)/3, (4 − )/3} = (2 ln 2)/3.

3
providedss of the value of the onstant .
we have have <0, that x< 1 k . Simil rly, for , x>0

f (x) > 2x + kHowever,>0provided that x > − 2 k. By Theorem 1.11, there exists number with c

.
(c) = 0

that . 6 2 for all . is giveexatlystradition to the

′
statement with ′
c f (p) = 0 f ′(x) = 3x + 2 > 0 x
f (c) = 0 and f (c′) = 0 for some c′ 6=c. Hene is on number with c

.
f (c) = 0
1
2 Exerise Set 1.1
tob. P2(0.5)
f (0.5) nd an bound for upper |f (0.5) − P2(0.5)|

()
9. aFindUsethe seond Taylor polynomial for f (x )= x cos x about , and:
x 0 =0

Findtheatualbounderrortofor.approximatetheerror , , and ompare this

. Approximate |f (x) − P2(x)| , for x in [0, 1].

1 1
d.SOLUTION:Findanupp Z
0 f (x) dx
Z
0 P2 (x
dx
Sinerbound for theusingerror in part . .

we have x
f′(x) = e (cos x − sin x),
f ′′(x) = −2e (sin x),
x and f ′′′(x) = −2e (sin x + cos x),
x
f (0 ) = 1 , , and So
′
f (0 )= 1 f ′′ (0 )= 0 .

[0 the endpoints .Thisimpliofthesinterval,thatthe and and minimum of

,0 .5] max imu m values x
e (sin x + cos x ) on our at
[0 , 0 .5]

ξ
a. We have P2(x) = 1 + and R2(x) = −2 (sin ξ + cos ξ) x3.
3!

P2(0.5) = 1 + 0.5 = 1.5 and

ξ |
|f (0.5) − P2(0.5)| ≤ ξ∈[0.0.5] − 2e (sin ξ + cos ξ) ≤ 1 ξ∈[0,0.5] |
To maximize this quantity on max
3!

(0.5)2 3

(0.5)2 max eξ(sin ξ + c s ξ) .

[0 , 0 .5]
, rst note that Dx x x
, for all in
(sin x + cos x) = 2e cos x > 0 x

Hene e0(sin 0 + cos 0) = 1 < e0 5(sin 0.5 + cos 0.5) ≈ 2.24.

.

1
all
3

b. A similar analysis to that|fin(0.part5)−(a)P gives,(0.5)|for≤ 2 3 (0.5) (2.24) ≈ 0.0932.

x ∈ [0, 1],
1
31
. |f (x) − P2(x)| ≤ 3 (1.0) e (sin 1 + cos 1) ≈ 1.252.

1 1 x2 1
3
Z
d. From part (b), Z0 f (x) dx ≈ 0 1 + x dx = x+ 2 0 = 2 .

error |1.378 − 1.5| ≈ 0.12.

1 1 1 1

Z Z Z
Sine 0 |R2(x)| dx ≤ 0 3
e1(cos 1 + sin 1)x3 dx = 0 1.252x3 dx = 0.313.

the at ualZ0
1 x 1 = 2 (cos 1 + sin 1) − 2 (1 + 0) ≈ 1.378,
x iscos x dx = e2 (c os x + sin x ) 0

e 1
Mathematial Preliminaries
14. approximateUsetheerror term of a Taylor polynomial to estimate the error involved in using sin x x to
3 radians. W e have

SOLUTION: First.we need to onvert the degree measure for the sine funtion to ≈

sin 1◦

180◦ = π
given by
radians, so 1 = we have
◦ π
180
radians. Sine f (x) = sin x, f (x) = cos , f (x) =
, and
′
The
′′
− sin x, and is

f ′′′(x) cos x, f (0) = 0 f ′(0) = 1, f ′′(0) = 0. approximation sin x x

− ≈
=
COS ξ 3

f (x) ≈ P2(x) and R2(x) = − 3! x . If we use the bound | cos ξ| ≤ 1, then

16. Let sin π − π = R2 π = − cos ξ π 3

x/
f ( ) = e 2 sin x
.
a Use Maple to determi 3
e third Malaurin polynomial

b. Find P3(x) .
(4)
SOLUTION: a .andDeboundne the error f (x ) |f (x ) − P 3 (x )| on . [0, 1]

f (x) by
x sin x
f := exp 2 the three term s of the Taylor s eries with
3
Then nd
rst
· (1/2)x 1

f := sin 3 x
g := taylor(f, x = 0, 4)
Ex trat the third M alaurin polynomial with
g := 1 3x +
1
6x2
+ 23 648 x 3 + O x 4

p3 := onvert polynom (g, )

b. Determine the fourth derivative. 1 1 2 23 3

p3 := 3 x+ 6x + 648 x

f 4 := diff (f, x, x, x, x)
fourth points .
[0, 1]

54 E
F ind th e fth derivative.

f 4 := − 1296E(1/ 2x ) 3x + (1/ 2x) cos 3x

119 1 5 1

f 5 := diff (
4, x)
See if the derivative has any ritial in 3x
5 := − 2592 E(1/ 2x) sin 3 x + 3888 E(1/ 2x ) cos

199 1 61 1

p := fsolve (f 5 = 0, x, 0..1)

The extreme values of the fourth derivative will ourat p := .604 7389 076

x= 0 1 , , or p.
4 Exerise Set 1.1

c1 := ( (x = p, f 4))
c1 := .09787176213
c2 := ( (x = 0, f 4))
c2 := .09259259259
c3 := evalf(subs(x = 1, f 4))
The maximum absolute value of c3 := 09472344463
funtion

error (4)
(x) is c1 and the error is given by
:= c1/24 error
24. In Example 3 it is stated that := .004077990089

a Show tha for all x we have | sin x| ≤ |x|. Use the following to verify this statement.
f (x ) is non- forallvaluesfuntion,ofand in partiular that
dereasing x

x ≥ 0 the funtion f (x) = x − sin x is non-dereasing, whih implies that

SOLUTION:b.sinUsexthe≤xfatFirstwiththatobserveequalitythesinethatonlyforwhen isx odd=0.to reah the onlusion.

f (x) = x − x we h ve f ′(x) = 1 − cos x ≥ 0 , beause

−1 ≤ cos x ≤ 1 for all values of x. Also, the statement le rly holds when |x| ≥ π, beause
|asin.Thex| observation≤1. implies that

number,and thatξexists between x1 and x2 with

f (x) > f (0) = 0 when x > 0. Hene for x ≥ 0, we have x ≥ sin x and when 0 ≤ x ≤ π, we have
|bsin.Whenx|= sin x ≤ x = |x|.
that −π < x < 0, we have π ≥ −x > 0. Sine sin x is an odd the fat (from part (a))

As asin(onsequene,−x)≤(−xfor)impliesallreal thatnumbers|sinx| = − sin x ≤ −x = |x|.

28. Suppose x we have | sin x| ≤ |x|.

a.
Showf∈thatC[a, ] 1 and x2 are in [a, b].

have opposite signs with (b) c1 and c

2

b. Suppose that f (ξ) = f (x1) + f (x2)

= 1 f (x1) + 1 f (x ).
2
2 2 2

with c1 and c are positive onstants. Show that a number ξ exists between x and x
2 1 2

c1f (x1 + c2f (x2)

. Give an example to show that the result in part
f (ξ)= does not neessarily . hold when

c1 + c2

aSOLUTION:.Thenumber c1 6=−c2 .

1
2 (f (x1) + f (x2))
Mathematial Preliminaries
Theoremistheaverage1.11ofthere f(x 1)
existandnumber,so it lies between these two values of
f (x2)

ξ between x1 and x2 with

f
5
. By the Intermediate Value

1 1 1
b. Let f (ξ) = 2 (f (x1) + f (x2)) = 2 f (x1) + 2 f (x2).

m = min{f (x1), f (x2)} and M = max{f (x ), f (x )}. Then m ≤ f (x ) ≤ M and 1 2 1

m ≤ f (x 2) ≤ M , so
ξ between x1 and x2 for whih

Thus c1m ≤ c1 (x1) ≤ c1M and c2m ≤ c2f (x2) ≤ c2M.

and (c1 + c2)m ≤ c1f (x1) + c2f (x2) ≤ (c1 + c2)M

c1f (x1) + c2f (x2)
applied to the inter val with endpoints

By the Intermediate Value Theoremm1≤.11 c1 + c2 ≤ M.

a number x1 and x2, there exists

. Let f (ξ) = c1f (x1) + c2f (x2) .

c1 + c2

f (x) = x ,
+ 1 x1 = 0 x2 = 1 c1 = 2 , , , and c = −1. Then f (x) > 0 for all values of x, but
2

c1f (x1) + c2f (x2) 2(1) − 1(2)

Exerise2..Find Setthelargest1.2,intervalpagein28 whih = = 0.
c1 + c2 2−1

.
SOLUTION: We need p∗ must lie to approximate 2 with relative error at most 10 √
−4

5. e. Use three-digitp √ √
p∗√−2
∗
omp√u2(0te .9999), 2(1.0001) .
that is, 2 ≤ 10−4, so p∗ − √2 ≤ √2 × 10−4;

This implies that √ 4 √ √ 4

, and
= 0.857 e = 2.72. So
− 2 × 10− ≤
arithmeti 13 = 0.9 29, 6

∗ − 2 ≤ 2 × 10− .
must be in the interval

rounding arithmeti to

13 6
SOLUTION:anddetermineUsingtheabsolutethree-digitandrelativeroundingerrors. 14 − gives 7,

2 − 5.4

14 7

13 6
14 − 7 = 0.07 20 and 2e − 5.4 = 5.44 − 5.40 = 0.0400.
1
6 Hene Exerise Set 1.2
3 6
The orret value is approximately 1. 14 − 7 = 0.0720 = 1.80.
954, so the absolute and relative errors to three digits are

2e − 5.4 0.0400

|
1.80 1.954 = 0.154 | 1.80 − 1.954 = 0.0788,
7. respetivelySOLUTION:.RepeatExerise.Using5(e)threeusing-diitthreehopping-digithoppingarithmetiandarithmetigives.

| − | 1.954

,
13
14 7
= 0.928
6
= 0.857 , and So e = 2.71.

1
3 6
Hene 14 − 7 = 0.07 0 and 2 − 5.4 = 5.42 − 5.40 = 0.0200.

| − | 13 6
onsiderably1.954
The orret value is approximately 1. 14 − 7 = 0.0710 = 3.55.
954, so the absolute and relative errors to three digits are

2 − 5.4 0.0200

arespetiv.Usethely.rstThethreeresultsterms in ofExerisethe Malaurin5(e) were 3. 55 1 .9 5 4 =1 .6 0 ,

andfor | 3.55 − 1.954 | = 0.817,
9. seri the artangentbetter. funtion to approximate

SOLUTION: Let P (x) = x 2

, and determine the absolute and relative errors.
1 x 3 + 1 x 5. Then P 1 = 0.46 4583 and P 1
= 0.3218107, so
π = 4 arc an
1
+ arctan 1
3

− 3 5 1 2 1 3
The absolute and relative errors are, respeti
π= 4 2
arctan + arctan 3 ≈ 3.145576.

v ely,

ex pon
12. Let | −
π 3.145576
|≈
3.983
×
10−3
and |π − 3.145576|
π ≈
1.268
×
10−3.

| |
x −x
a Find f (x) = e − e .
x

b. Use threelim-digitf(rounding). arithmeti to evaluate

x→0
SOLUTION:Replaeah. Sinential funtion with its third Malaurin . f (0.1 )
polynomial and repeat part (b).

Rule to giv limx→0 ex − e−x = 1 − 1 = 0 and limx→0 x = 0 , we an use L'Hospitals

lim ex − e−x = lim ex + e−x = 1+1
 = 2.
x→0 x x→0 1 1
Mathematial Preliminaries 7

b.With three-digit rounding arithmeti we have e 0.100 = 1.11 and e −0.100 = 0.905 , so
. The third Malaurin polynomialsf(0.100)give= 1.11 − 0.905 = 0.205 = 2.05.
0.100 0.100

1
2
1 and 1
2
1
3
so ex ≈ 1 + x + 2 x + 6 x3 −x ≈1−x+ 2 x − 6x ,
mahine
f (x) ≈
1+x+ 1 x2 + 1 x3
x
1 x + 1 x2 1 x3
= 2x +
x
1 x3
= 2 +3x.
3 1 2

Thus, with three- digit rounding, 2

we have−
6
− 2 −6

15. . Find t he deimal f(0.100 )equivale≈2 +

1
3
2
(0.100) = 2 + (0.333)(0.001) = 2.00 + 0.000333 = 2.00.

nt of the oating-point number

SOLUTION:001111111111Thisbinary mahine0101001100000000000000000000000000000000000000000000 numberisthedeimalnumber .

+ 21023−1023 1+ 2 + 2 + 2 +2 !
1 2 1 4 1 7 1 8
+
deimal= 2 1 + 4 + 16 + 128 256 = 1 + 256 = 1.32421875.
0 1 1 1 1 83
16. . Find the equivalents of the next lar gest and next sma llest oating-point mahine number to

SOLUTION: Thenextsmallest 00111 111 111 1 mahinenumberis

0101001 100 000 000 0000 000 000 000 000 000 000 0000 000 000 000 0 .

0 01111111111 0101001011111111111111111111111111111111111111111111

=1.32421875 − 21023−1023 2−52

and next largest mahinenumberis
= 1.324 218 7499 999 997 779 553 950 749 6869 191 527 366 638 183 5937 5 ,

0 01111111111 0101001100000000000000000000000000000000000000000001 =1.32421875 + 21023−1023

2−52
21. a. Show that the =1 .32 421 875 000 000 0222 044 604 925 031 308 084 7263 336 181 640 625 polynomialnestingtehniqueanbeusedtoevaluate .

f (x) = 1.01e4x − 4.62e3x − 3.11e2x + 12.2ex − 1.99.

 and three-digit rounding

8 Exerise Set 1.2

b. Use three-digit ari hmeti and the formula given in the statement of part ( ) to evaluate
. Redo .the alulatiapproximationsroundingpart(b) using the
f (1 .53 )
esting form of

SOLUTION:d.Comparethe. Sine in parts (b)annd (). f (x) that was found in part (a).

, we
enx = (ex)
n
write
b. Using f (x) = ((((1.01)ex − 4.62) x − 3.11) ex + 12.2) ex − 1.99.

1.53
= 4.62 gives e2(1.53) = (4.62)2 = 21.3 ,
e3(1.53) = (4.62)2(4.62) = (21.3)(4.62) = 98.4, e4(1.53) = (98.4)(4.62) = 455
= 1.01(455) − 4.62(98.4) − 3.11(21.3) + 12.2(4.62) − 1.99
and . So f (1.53)

= 460 − 455 − 66.2 + 56.4 − 1.99

= 5.00 − 66.2 + 56.4 − 1.99

. We have
= −61.2 + 56.4 − 1.99 = −4.80 − 1.99 = −6.79.

f (1.53) = (((1.01)4.62 − 4.62)4.62 − 3.11)4.62 + 12.2)4.62 − 1.99

= (((4.67 − 4.62)4.62 − 3.11)4.62 + 12.2)4.62 − 1.99
= ((0.231 − 3.11)4.62 + 12 2)4.62 − 1.99
d. The exat result is = (−13.3 + 12.2)4.62 − 1.99 = −7.07.

7.61 , so the absolute errors in parts (b) and () are, respetively,

| − 6.79 + 7.61| = 0.82 and | − 7.07 + 7.61| = 0.54. The relative errors are, respetively, 0.108 and
24. Suppose that
0 .0 710.

f l(y) is a k-digit rounding approximation to y. Show that

y f (y) k+1
SOLUTION: We will onsider the −y in
≤ 0.5 × 10− .
two ases, rst when
When solution
, and then when
dk+1 ≤ 5

dk+1 > 5.
dk+1 ≤ 5, we have
When y f (y) = 0.dk+1 . . . × 10
n−k
0.5 × 10−k = 0.5 10−k+1.
−y 0.d1 ... ×
10n≤ 0.1 ×

dk+1 > 5, we have

f() (1 0.dk+1 . . .) 10n−k (1 0.5) 10−k

Hene the = in all situations. < − ×
= 0.5 10−k+1.
−
×

− 0.d1 . . . ×
10n 0.1 ×

inequality holds
Mathematial Preliminaries
28. SOLUTION:areShowonsisthatentbothwithForsetsthetheofinitialideatagdata,ivenslawwein.thehaveopening appliation for this hapter an give values of that
9 T

0.995 ≤ P ≤ 1.005, 0.0995 ≤ V ≤ 0.1005,

This implies that 0.082055 ≤ R ≤ 0.082065, and 0.004195 ≤ N ≤ 0.004205.

Sine 287.61 ≤ T ≤ 293.42.

15◦ Celsius = 288.16 kelvin, we are within the bound. When P is doubled and V is halved,
so 1.99 ≤ P ≤ 2.01 and 0.0497 ≤ V ≤ 0.0503,

Sine ◦
atual19 292.16
286.61 ≤ T ≤ 293.72.
temperatCelsiures=is kelvin,we are again within the bound. In either ase it is possible that the

Exerise3.a.DetermineSet1the.3,numberpage 290.1539 = 17◦ Celsius.

SOLUTION: a. Sin the terms f the

ensure 10− needed π
n of terms of the series

2i−1
∞ i+1 x
X −

arctan x = lim P (x) = ( 1)

that are required to ensure that →∞ n =1 − (2i 1)

to the . series
b. How many terms are required 4P (1) π < 10−3

| − |
n
series
10 auray for an approximation to ?

∞ 1
Xi −
1)i+1
thealternatenextterminsign,.Totheensureerrorsigniproduedantauray,bytrunatingweneedthe
π = 4 arctan 1 = 4 (
− 2i 1
to hooseatany term is less than the magnitude of

=1

n so that
onvergent approximation
4 or
−3
So 2(n + 1) − 1 < 10 4000 < 2 + 1.
b. In this≥2000ase,. we need

4 10 or
Clearly, a more rapidly method is needed for this
< 10 − n .
> 20,00 0,000,0 00 .

2( + 1) − 1
10 Exerise Set 1.3

5.
so that the X
Another formula for omputing π an be dedued from the identity

π−
m
1 X
ed to ensur
1
e an

.
determine
approximation to
− i
Determine the number of terms that4must=4beactansum 5 − arctan 239
.
SOLUTION: The identity implies that
π to within 10−3

∞ 1 ∞ 1
alulations

The seond sum is muh =1

π= 4
−
( 1)
i+1
5 2− 1 (2 1) −
i=1 −
( 1)
i+1
239 2 − 1 (2i 1)
smaller han the r t sum. So we need to the minimal value of

. We have
st term of the rst sum is less than 10
i+ 1 −3

for a total of omputed,itheinner sum X X number

−1 additions,
4 4 4 4 4 4 4

, = = = 2.56 × 10−
aSo.How3iter:= 1 : 51(1) = 5 3
i = 2 : 5 (3) 375 and =3: 55(5) 15625 .

8. msanyare sufient. are needed to a sum of the form

One the = 2 multipliations and

n i

b.sumSOLUTION:Re.-express thea. Forserieseahin a way that will redue theai j? of alulations needed to determine this
n inner − 1 additions required
=1 j=1

b. By thesumssumareas multipliations i and − additions.

rewriting
n(n + 1)
ultipliations and n
n(n + 1)
Xi X
=1 i=1
The nal total is: alulation are , wefornowthe neednalsum.i−1 additions to sum
 j signi antly

weOneanthe redue the amount−1of= 2

. For eah .
− n additions

(n + 1) (n + 2)(n 1)
2 2
ipliations

Theprodutstotal. bj 1 n mult ai −1

been'sforreduetotal of needis 2 (n + 2)(n − 1)

's, followed by multipliations

X
i i n i
X X X

bj = ai bj
=1 j=1 i=1 j=1
X
i (n + 1)
=1
ad'sitiarenssummed,bythismethwe still by the additions of the

d from
1 , but the number of has
2 n(n + 1) o n.
Mathematial Preliminaries 11

10. Devise an algorithm to ompute the real roots of a quadrati quation in the most efient manner.
SOLUTION: The following algorithm uses the most effetive formula for omputing the roots of a
quadrati equation.
INPUTOUTPUT , . A BC

Step 1 If x1, x2.

A = 0 then if
B = 0 elsethen set NO SOLUTIONS');

OUTPUT (`ONE SOLUTION',

x1 = −C/B

2 Set STOP. x1);

Step 3 If D = B − 4AC. 2

OUTPUT (`MULTIPLE; ROOTS',

then set = −B/(2A)
);
D=0 x 1

Step 4 If STOP.
x1

D<0 then set √ ;

b= −D/(2A)
a = −B/(`COMPLEX(2A); CONJUGATE ROOTS');

x1 = a + bi

OUTPUT ; x2 = a − bi

Step 5 If STOP. (x1, x2);

B≥0 then set √ ;

d=B+ D
x1 = −2C/d;
else set x2 = −d/(2A)
d = −B +
√ ;
D

x1 = d/(2A);
Step 6 OUTPUT x2 = 2C/d.
15. SupposeSTOPthatas. (x1, x2);

x approahes zero,
α β
Let F1(x) = L1 + O (x ) and F2(x) = L2 + O x .
c1 and c2 be nonzero onstants, and dene

Show that if F (x) = c1F1(x) + c2F2(x) and G(x) = F1(c1x) + F2(c2x).

a. γ= minimum , then as approahes zero, {α, β} x

γ
F (x) = c1L1 + c2L2 + O (x )
12 SOLUTION: |x | we have positive o nsta nts k 1 and k 2 independent of
Exerise Set 1.3
γ
Suppose for sufiently small

b. G(x ) = 1 + L2 + O (x )

x , for whih
α β
Let |F1(x) − L1| ≤ K1|x| |F2(x) − L2| ≤ K2|x| .
a. Wec=havemax (|c1|, |c2|, 1), K = max (K1, K2), and δ = max (α, β).

|F (x) − c1L1 − c2L2| =|c1(F1(x) − L1) + c2(F2(x) − L2)|

α β
≤|c1|K1|x| + |c2|K2|x|
α β
cK x + x|

≤ || | γ
for sufi ntly small cK x γ 1 + x δ−γ Kx ,

b. We have . Thus, ≤ || ||
γ . ≤ ||
|x| F (x) = c1L1 + c2L2 + O (x )

|G(x) − L1 − L2| =|F1(c1x) + F2(c2x) − L1 − L2|

α β
≤K1|c1x| + K2|c2x|
δ α β
Sine Kc x + x|

≤ || | γ
for sufiently small Kcδ xγ 1+ x δ−γ K′′ x ,

≤ | | γ | | ≤ | |
16.

Dene
Consider the Fibonai .sequeneThus, de ned |x | G( x )= L 1
by

+ L2 + O (x )
.

F0 = 1, F1 onverges,=1andFn+2 = Fn+1 + Fn
, if . n ≥ 0
is the golden
ratio: x n = F +1/F n . Assuming that n→∞ x n = x
lim
show that the limit
SOLUTION: x =1 + √
5 .
/2

17. Fibonai sequene also satises the solution equation

x= 1 + √5
/2
1
.
we have lim xn = lim x +1 = x and x =1+
n+1 ,

→∞ n→∞ xn
1
qu adratiwhih impliesisthat
2

Theonly posit ve x to= this1+ x , x − x − 1 = 0.

√
a riteM apleM aple
Fn ≡ Fn = 5 " √ ! 2
n − −2
√ ! n#.
1 1+ 5 1 5
proedure to alulate˜
b. Use with the default value of F100
Digits.followed by evalf to alulate

ed.. Whathyisresults ˜ F100

ompute(b )? art.(a)?

fromevalfresultparttofrom
the
the resultwhenfrompartyouuse(a)(b)themoreobtainedommandauratemoresimplifythanrapidlytheinsteadthanresultof

˜
F100?
Mathematial Preliminaries termsstepinofonethelinesequene.Mapleare:would13
lineswewill.TheshowproeduretheMaplefor outpalulatingforeahthe

SOLUTION:produethisoutputa.Toonsaveparatspae

n := 98; f := 1; s := 1
for n := 98 f := 1 s := 1

i from to do 1 n

l := f + s; f := s; s := l; od :

l :=2 f := 1 s := 2
l :=3 . f := 2 s := 3

l :=218922995834555169026 f := 135301852344706746049 s := 218922995834555169026

b. Wel have:=354224848179261915075

sqrt sqrt 1 (1 +(5)

100
1−
sqrt (5)
100

F 100 :=
evalf
(5)
F 100 := √ 5
2 −2 +2
2
√5 −2
! −2 √5 !
100 100
1 1 1 1 1

(F 100)
de..omput.TheTherresultdsingispartt het enpart- sam e(a)digit(a) sreqasroomputedthendingiredresulttr aversi ngusi ngrithm etipart exat(a) .loop.int eg er98tim esa rithmeti,.21 and t he result in part (b) is

0.3542248538 × 10
14 Exerise Set 1.3
Solutions of Equations of

One Variable Exerise Set 2.1, page 54

1. Use the Bisetion method to nd p for f (x) = 3

√
x − cos x on [0, 1]
1 = 0 and b1 = 1, so f (a1) = −1 and
f (b1) = 0.45970. We have

SOLUTION: Using the Bisetion method gives

1 1 .
Sine p1 = 2 (a1 + b1) = 2 and f (p1) = −0.17048 < 0.
f (a1) < 0 and f (p1) < 0, we assign a2 = p1 = 0.5 and b2 = b1 = 1. Thus

1
Sine f (a2) = −0.17048 < 0, f (b2) = 0.45970 > 0, and p2 = 2 ( 2 + b2) = 0.75.

f (p2) = 0.13434 > 0 , we have a 3 = 0.5 b3 = p3 ; = 0.75 so that

1
2. a. Let p3 = 2 (a3 + b3) = 0.625.

SOLUTION: Sine 1 . Use the Bisetion method on the to nd .

f (x) = 3(x + 1) x − 2 (x − 1) 1 interval [−2, 1.5] p3

we have the following sign graph for f (x ) = 3(x + 1)

x− 2 (x − 1),
:
Thus,assign a = −2, with f (a ) < 0, and b
f (x)

2 2 2 = − 4 , with f (b2) > 0. Thus, p2 = −1.125 and

x1 1 2 22222 0 11 111 1 11111111111 1
f (p2) < 0. we assig n a 3 = p 2 = 15 and
−1 .12 5 3
b = −0. 25 . Then 3
p = −0. 6875 .
x2 1
2 2 2 2 2 2 2 2 2 2 2 2 0 11 1 1 1 1 11 1 1 1 1
2
2

x2 1 22222222222222 01 1 1 1 11 1 1 1 1
f
(x) 2 2 2 2 2 2 0 1 1 111 0 2 01 1 1 1 11 1 1 1 1
22
x
1
21 0 2 1 2 3

with . Sine 1 , we have .

We 1= Hene, with
−2 f ( , and ,
1) < 0 1 = 1. 5
b
1 f (b1) > 0 p1 = − 4 f (p1) > 0
16 approximation . to within to the rst positive value of
−5
10 x
Exerise Set 2.1
8. a Sk th the graphs of y = x and y = withb.Use the tan
Bisetion method to nd an
graphsSOLUTION:.Thexg=aphstanxof.

ross neary = x and y = tan x are shown in the gure. From the graph it appears that the
x = 4.5.

y
10 Y=X

5
x
10
b. Beause Y = TAN X
210

g(x) = x − tan x has

the fat that (4.4) ≈ 1.303 > 0 and g(4.6) ≈ −4.260 < 0,

Using Alg rig ismontinuous2.1gives on [4.4, 4.6] gives us reasonable interval to start the bisetion proess.
.
11. Let
p16 = 4.4934143, whih is au ate to within 10−5
for the following intervals? 3 . To whih zero of does the Bisetion method

onvergea.f(x) = (x + 2)(x + 1)x(x − 1) (x − 2) f

.SOLUTION:
[−3, 2.5]
[−1.75,1.5]Sine

we have the following sign graphf(x)for= (x + 2)(x + 1)x(x − 1) (x − 2), 3

f (x).

x 1 22 2 2 2 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
x11 22222222 0 111111111111111111
x 222222222222 0 11111111111111
)
a. The interval 2 3
(x 1 2222222222222222 0 1111111111
x2 2 22222222222222222222 0 111111
f
(x) 22220 111 0 222 0 111 0222 0 111111

23 22 21 0 1 2 3 x

[−3, 2.5] ontains all 5 zeros of . F r f a1 = −3 , with , and

f (a1) < 0 b1 = 2.5 , with
with
f (b1) > 0
, we have
p1 = (−3 + 2.5)/2 = −0.25 , so
f (p1) < 0 . Thus we assign
a2 = p1 = −0.25 ,

f (a , and , with .
2) < 0 2 = b 1 = 2.5
b f (b 1) > 0
Solutions of Equations of One Variable

12. Usethe Bisetion Algorithm ,thetomethodnd

ontains

b1 = 1.5 , with , we have f (b 1) < 0 p 1

−1
[−1.75, 1.5] ontains the zeros onverges to to. −1
−1

.The n we assign
= (− 1.75 + 1.5 ) /2 =
1

−0.
= −1.75

125 and f ( 1)< 0

with f (a1)

.
>0 , and
17

. Then we assign

SOLUTION: The dened by approximation √3 that is aurate to within 10−4

andHene p 2
= (−0.25 + 2.5)/2 = 1.125 and f (p2) < 0
3
= 1.125 , with f( 3) <0 ,
Bi tion method ,towiththis funtion on the i terval 2 has √ as its only zero. Applying the

The interval,with funtion . Sine f (x) = x −3 3 positive

. b3 = 2.5 f (b3) > 0 [1.125, 2.5]

starting derease , 0,iteration1. s
2 = 1 =only−1the.75zero ( 1) > 0, and 2= 1 = −0.125, with f (b2) < 0. Sine [−1.75, −0.125]

14. equationUseTheoreminterval2.would1tond boundthefornumberthe giveshat √

approximation
numberofiterationsof areneededrequiredtoapproximate..Usingasolutionasmallerto the

[1, 2] 3≈ 14 = 1.7320

gives the sequene. bound

n ≥ 12 . Applying the Bisetion Algorithm
SOLUTION:isneededwithi nx the +intervalFirstx 4not e=a nd 0thato nthethea urayint epa rtiularvalreq uirement[1,4 ]
3
equationtoa n.aurayplaysTo noda npa rtof in nding t het hatbo und;isa urat eallthattois
10− 3 .

10−3 , we need to det rmine the number of iterations n so that

As a onsequene, a p p for b < − a = 4−1 < 0.001; 3 10 3 <2.
the n umber of iterations is that is,

| − | 2
n
2
n
×
divergent armoni armon onvergent
However, example
17. Dene
L = 10 |pin−
12

|
p = 1.37 87

19. suspetAtualtroughsolu.of div erg ene

n−1
and
sequene {p } by p = k
1 . Show that lim (p − p
Integral
, even though the
−1) = 0

X
partialwhoseofThenalyzethepopproxsummanttheConsetermsfofpartialthegoolthetoprofo zero,sumnblemthistobuofseries,i snotrootathenotbasedseriestheanyrapidlyfatalulusonserienoughthenotherondition.thistextThepartiulshouldisproduebasedthgriveonsaseriesqueneattheleastisdiverges,thetwoserieslassi.TOnet.itproofThereisthatwillaretestofsimplyisathenyforriesproofs an

k=1 →∞

SOLUTION:{p }Sinediverges.

p − p −1 = 1/n , we have lim →∞(p − p −1) = 0. n th

 =1 foot. When lled with water to within a distane of the the volume h top, V = 12.4 ft
water feet as ross the shape of a s miirle with radius
utiveofaretermslengthseekingofa. sequene might be losetioneah other, butissmallnotsufshouldientalwayslosebeto the
a tionwateryou withi n 0.01 feet.

of the
is given by the formula 3

water to
1/ 2i
Determine the depth of the 12.4 = 10 0.5π − arcsin h − h 1 − h2
18 Exerise Set 2.2

SOLUTION: Applying the Bisetion Algorithm on the interval [0, 1] to the funtion

h 1/ 2i
gives f (h) = 12.4 − 10 0.5π − arcsin h − 1 − h2
Exerise3.Thef llowingSet2.methods2,page,areso64proposedthedepth tois ompute feet.
h ≈p 13 = 0.16 17 r − h ≈ 1 − 0.1617 = 0.8383

spe dof onvergene, assuming 211/3 . Rank in order, based on their apparent
them

a p0 = 1.

20pn−1 + 21/pn2−1
2
b. p = n
1
p3 −1 − 21
. pn = pn−1 − 3pn2−1
pn4−1 − 21pn−1

d. pn = pn−1 − p2 −1 − 21
SOLUTION: a. 1/ 2

2
1

p= pn−1

Sine
20x + 21/x2 20 1
and pn = 20pn−1 + 21/pn2−1 , we have g(x) = = x+ 2 ,
21 21 21 x
20 2 20 2 6

b. Sineg′(x) =
21 − x3 . Thus, g ′
211/3 = 21 − 21 = 7 ≈ 0.857.

p =p − pn3−1 − 21 , we have g(x) = x − x3 − 21 = x− 1x + 7 =2 x + 7

and n−1 3pn2 −1 3x2 3 x2 3 x2

2 7 2 1 1

. Sineg′(x) =
3 −x3 . Thus, g′ 211/3 =
3 −3 = 3 .
= 0.33 3

we have pn 4 −1 − 21pn−1

pn = pn−1 − pn2−1 − 21 ,

g(x) = x x4 − 21x = x3 − 21x − x4 + 21x = x3 − x4

and
2
− x − 21 x2 − 21 x2 − 21

g′(x) = x2 − 21 3x2 − 4x3 − x3 − x4 2x 4 2 5 3 4

= 3x − 63x − 4x + 84x − 2x + 2x
5
2 21)2 (x2 21)2

(x − −
5 4 3 2
= −2x + x + 84x − 63x .

(x2 − 21)2
Solutions of Equations of One Variable
xed approximation √3 that is aurate to within
19
dThus.Sineg′ 21 1/3
≈ 5.706 > 1 .
21 1/2 21 1/2 √ 21
and p = pn−1 , we have g( ) = x = x1/2 ,
√
Realling onverg ene

′ − 21 . Thus, 1 .
9. TheUseo rdera of- 2x
()= ′ 211/3 = − 2
3/2

pointnvergeneiterationwouldmethodlikelytodeterminebe(b),(d),an( ). Choie () willto not likely onverge.

willSOLUTION:produe.therapidsolutionAsalwaysto Exerisewith.xed12-pointSetioniteration,2.1,thewetrikneedistotoonverthoosethetherootxed--ndingpoint problem that

10−4

positive solution of g and an appropriate terval that produes an approximation to a

f (x) = x2 −3 into xed-point problem. One suessful solution is to write

numerous 3
then add 0=x −3 2
as x= x ,
x to both sides of the latter equation and divide by 2. This gives g(x) = 0.5 + 3 , and for
12. . , we have
0 = 1.0 √3 ≈ p 4 = 1.73205
x

Determine xed-point funtion .

are possibilities: .
SOLUTION: There3x2 − x that is aurate to within
=0 −5
10

[4, 5]. So weneed a different xed-andpoint funtion. If we note that x = tan x implies that
For g (x ) = q 3 2 [0, 1] with p 0 , we
=1 p12 = 0.91 0015 .

1x

14. Use a xed-point iteration method to determine,havea solution aurate. to within g(x) = ln 3x [3, 4] p0 =4 p16 = 3.733090

obtain, p0 = 4: 10−4 for x = tan x, for

x SOLUTION:in . Using [45]

18. suf p3 p4 agree |g′ (x)| ≤ k is replaed by the statement “g (x) ≤ k < 1, for all ′

(x) = tan 0 = 4 gives p1 = g(p0) ≈ 1.158, whih is not in the interval

1 1 1 1
we again with x = tan x and dene g(x) = x + t x − x

Beause p1 ≈ 4.61 69, p2 = 4.49596, 3 = 4.49341 and p4 = 4.49341.

nd
iently aurate. 3istotruevedeimalif plaes t is re sonable to assume that these values are

b. Show that”. Theorem 2.4 may not hold when

x ∈ [a, b]
“
|g′(x)| ≤ k is replaed by the statement
g (x) ′ k<1 , for all x [a, b] ”.
≤ ∈
20
SOLUTION:inpoints a. The proof of existene is unhange
 bgiving.ForTheoremthesame 2ontradition.4,onsider as in Theorem 2 .3 .
pq = (p ) g () = g ′( ξ) (p
−
−

1
and.Useto
 19. p = 2

onverges for any

x
n−1

b. thatt if
in (a)
. Show
0< x 0<

OLUTION: a. First let

then

2
Suppose, as is the, soassumption given.Also,in

1/x
 where

x0 1

and
0

Thus,

p = limm→∞ xm. Then xm−1

Thus
so
2 p

b. Consider

p= ±
0 < x0 < 2

so

0
2
Solutions of Equations of One Variable
that if , whih imphaslmpliesathat -p
g x ed int at , that
p g , and that
C[a, b ] g ′
21
exists in . Show (a, b )

. To 1:omplete the problem, we onsider the three possibilities for x 0 .

>0

2 √ by part (a) implies that √ .

2 x0 > limm→∞ xm = 2
Case 3 x0 =√ 2 m = √ 2
for all m and that m→∞ x m =
lim
√ 2
.

SOLUTION:number √ , whih i that √

√ √
Sinefor g so me′isontinuousvalue of ′ ′
.at p and |g (p)| > 1, b y letting ǫ = |g (p)| − 1 there ex ists a

In any ituation, the

0 < x0 < 2 and
2 < x1 by part (b). Thus
sequene onverges to < . . . < x1 im xm = 2.
0< x 0 < 2< x m+1 < x m

24. Sup os e t hat t he funtion

√ m→∞ , and r pidly, as we will disoveri nitial he S etion 2.3 .

exept g′(p) >1 ∈ p

|
if
| , then the xed-point sequene will fail to onverge for any hoie of 0 ,
, we have
x satisf yin g 0 < |x − p| < δ

p =p

δ>0 suh that

whenever |g′(x) − g′(p)| < ε = |g′(p)| − 1,
0 < |x − p| < δ. Sine

for any |g ′(x) − g′(p)| ≥ |g′(p)| − |g′(x)|,

the equene move away fromapproximation gets p, but is not equal to p, the sueeding terms of

Exerise1.Let Set 2.3, page 75p. So the sequene annot onverge to . p

If |g′(x)| ≥ |g′ (p)| − |g′(x) − g′(p)| > |g′ (p)| − (|g′(p)| − 1) = 1.

p0 is hosen so that 0 < | − p 0| < δ , we have by the Mean Value Theorem that
for some |p − p| = | 1 (p0) − g(p)| = | ′
(ξ)||p0 − p|,

between and Thus, ξ 0 . 0 < |p − ξ| < δ and

This mea s that when ′ lose(ξ)pto p >p p .
| 1 − p| = | || 0 − | | 0− |

SOLUTION: Let
2

f (x) = x − 6
and
p0 = 1
. Use Newton's method to nd
.
p2

f (x) = x2 − 6 . Then f (x) = 2x, and Newton's method beomes

′

pn2 −1
−6 .
With p =p f (pn−1) =p
n n−1 −
f ′(pn 1) n−1 − 2pn 1
− −
p0 = 1, we have

and p =p p0 2 − 6 = 1 1−6 = 1 + 2.5 = 3.5

1 0 − 2p0 −2

p =p p12 − 6 = 3.5 3.52 − 6 = 2.60714.

2 1 − 2p1 − 2(3.5)
22 Exerise Set 2.3
3. LetSOLUTION:F alse.WhihPositionmethod.Theg ivesfor mulab etterf orre sults?b oththe Se ant methodf orand(a)thethe method Seant ofm ethodF alse Positionand(b )theis meth od of
2 . With and , nd
f (x) = x − 6 p0 =3 p1 = 2 p3

a.WithThe Seant method: p =p f (pn−1)(p −1 − pn−2) .

n n−1 − f (pn−1) − f (pn−2)

gives and , we have p0 = 3 p1 = 2 f (p0) = 9 − 6 = 3 and f (p1) = 4 − 6 = −2 . The Seant method

and p =p f (p1)(p1 − p0) =2 (−2)(2 − 3) = 2 2 = 2.4
−
2 1 f (p1) − f (p0) − −2 − 3 − −5
f (p2) = 2.42 − 6 = −0.24 . Then we have
f (p2)(p2
Position:
p1) ( 0.24)(2.4 2) 0.096
−
− − −

Withb.The method of False p=p = 2.4 = 2.4 = 2.45454.

−
3 2
f (p2) − f (p1) − (−0.24 − (−2) − 1.76

0 =3 and p = 2, we have f (p ) = 3 and f (p ) = −2. As in the Seant m th d (part (a)),

1 0 1

pa2reassignment=2.4andf (ofp2) = −0.24. Sine f (p1) < 0 and f (p2) < 0, the method of F lse Position requires
p1. Then p1 so that , with , and , with 0 p 1 =3 f (p 1)= 3 p 2 = 2.4

f ( 2) = −0.24. We alulate p3 by

Sine p f (p2)(p2 − p1) (−0.24)(2.4 − 3) 0.144

=p = 2.4 = 2.4 = 2.44444.
−
3 2
f (p2) − f (p1) − −0.24 − 3 − −3.24
ApplymationsNewton'swouldmethodshow,thetoauraythatndthesolutionSeantofhemethodto is better.is the same. Continuing to more

. √ is hanged toapproximations
5.
approx 6 ≈ 2.44949
Newton's
SOLUTION:
within
With. x − cos x = 0 in the interval [0, π/2] that is aurate
−4
10
f (x) = x cos x , we have ′
f (x ) = 1 + sin x

−
For =p pn−1 − cos p −1 .
n −1 − 1 + sin p −1

7. . Apply the,weSeanthavemethod ,to nd a 0 =0 p 1 =1p 2 = 0.75036

, top 3 = 0.73911
, and p4 = 0.73909.

to within solution x are

cos x = 0 in the interval [0, π/2] that is aurate
SOLUTION: The. Seant method ns − by the sequene
10−4

method is approximati generated

, and the sequene generated by
p =p (pn−1 − cos pn−1)(pn−1 − pn−2) .

n n−1 − (pn−1 − cos pn−1) − (pn−2 − cos pn−2)

Solutions of Equations of One Variable 23

Using the endpoints of the intervals as p0 and p1, we have the entries in the following table.
n pn

0 0
1 1.5707963
2 0.6110155
3 0.7232695
4 0.7395671
5 0.7390834
9. aurate.Apply the method of False Position to nd approximationssolution to 6 0.739085 1

SOLUTION:Ex erise7 ,in

to within b utTheinorporatesm ethod.ofF alsethe additiPositionnal b rak ting
x − cos x = 0
test.areUsinggeneratthee using thisof thesameintervals
ndpointstheinterval th t is [0, π/2]
form ulas as

−4
10

p0 and , we have the entries in the foll wing table.

p1

n pn

0 0
1 1.5707963
2 0.6110155
3 0.7232695
4 0.7372659
5 0.7388778
6 0.7390615
13. App y Newto 's method to nd a solution, aurate towithin 7 0.739 082 5

the losest point on the graph of 10−4 , to the value of that produes x

SOLUTION: The distane betweeny=anx2arbitrarytothepointpoint(1, 0).

point ois the graph method 0 = 1 gives p5 = 0.589755. The
on the graph of
(1, 0) x, x2 y = x2 and the point
Applyi g Newton's tothatndisthelosestroot toof this equationhasthe with
of y = x (1 ,0) approx imate oordinates
this funtion, q − − , it is easier to work with the square of
p

2 2 2 4 2
Bea se a derivative is needed to nd the ritial points of (x )= (x 1) +(x 0)= x +x − 2x + 1.
d

2 4 2
whose minimum will our at the same value of f( x )= [d(x )] = x +x − 2x + 1,
as the minimum of . To minimize we
need
′
(x) = 4x3 + 2x
x d(x ) f (x )

x so that 0=
2.
−
2
(0.589755, 0.347811).
24
16.
Exerise Set 2.3
Use Newton's method to solve for roots of

SOLUTION: Newton's method with0= 12 1

+ 4x
2 1
− x sin x − 2 cos 2x.

p0 = π p15 = 1.895488 and with p0 = 5π gives

19. 2 gives

ThefuntionwithequivExplresultssmallinlentandwhyformmagnitude,othenotderivative.Withiterationindiatewhihaveequationinreasesfastthe,theonvsameforsequenethergenerootsSeantround.doesAu-offmethoduallywenoterrorapprassonverge.shouldiatedahnotrinwiot,200behweNewton'susediterationsareindividingthemethod.algebraiallybybeausenumbersthe

p19 = 1.895489 p0 = 10π

f (p 1)pn 2 f (p 2)pn 1
SOLUTION:andThisdenominatorformulanumerator when =
the
−
− − ofnearly equal nu bers in bo
− − .
the

f (p −1) − f (p −2)

subtration
22. AlgoUsepMapleferrithmthee
sub tratsto approahdeter mineo rretion.how manyf ro miterati onsresult ofthat Newton' s areshouldnea rlydo minmethodeq u

p −1 and pn−2
form g ive

alte.withThethealulations t
in.This

e Seantisalways

inorporates
root of p 0 = π /4 are needed tond
SOLUTION: We rst dene .
f (x) = cos x x to within −100
10

−
f (x) and f (x) with
′

f := x− > os(x) − x
and f := x → cos(x) − x

forp := x− > (D)(f )(x)

We wish to use 100-digit rounding
arithmetifp:=xso→we−setsin(x) − 1

Digits := 100; p0 := P i/4 Digits

:= 100
1
n from to do 1 7
p0 := 4π

p1 := evalf(p0 − f (p0)/f p(p0))

err := abs(p1 − p0)
p0 := p1
odThis gives

p7 = .73908513321516064165531208767387340401341175890075746496 whih is
aurate to 56806357732846548835475945993761069317665319,
10−100 .
Solutions of Equations of One Variable 25
approximation 2 within
th smallest positive
0.4x
zero.has an innite numb er of zeros.

23. aThe funtion d nedonlybyfnegative()=lnzerox to+ 1 − c πx

b. We an initial for the

.
. Find an p0 = −0.5 suf iently p3 = −0.4341431.

te the four smallestpositivezeros to within.

Approximat
b. e approximations 10−6
10−6

SOLUTION: The key to thisdominatedproblis reognizing the beha .ior of

det. 25 h to within 10−6

erm goes to zero, so 0.4x . When is negative, his

x

he alulations, and ()
by ln 2
+1 . However, when x
0.4x dominates
wh f (x) will be z ro approximately
p0 = 24.5 . A aurate
suf iently approximation ; to the 25th smallestis positivepsitive, zero is
when integthis m makes no o tribution; tha is,
 a
Graphs
rminecosπxinitial=0. This ours whenprodx =esn/the2 llowingpositiveresults: er n. Using this nformation to ; andparts to g ive

use: to nd the aurate

p0 = 0.5 to g ve p3 = 0.4506567 0 = 1.5 to give p3 = 1.7447381; p0 = 2.5

reasonable initialtoappgiveoximation for the
. In general, a
5 = 2.23 831 98
0 = 3.5 p4 = 3.7090412.

d. Let nth positive root is n − 0.5.

2= 24 for.4998870various. of the region are shown below.

y y y
12 600 30000
400 20000
8
200 10000
4
26. Determine the inimal annual interest rate 12 16 20
4 6 x x
24 22 2 x 4 8 16 210000
2200
24 220000

invested to aumulate an amount i at whih an am unt P = $1500 per month an be

equation A = $750, 000 at the end of 20 years based on the annuity due
The approximate is b itandumb ersosolutionha vemt onforthis adjustedthiseq uationprob leb eausean,sineb ethefoundpay mthed b yrivatnts any arev em adeof monthlyinthisserathe rtion thanNewton' sy arly.
hoie interest method
P
roblem where the funtion is given by

SOLUTION: This is simply a root-ndingA=p i [(1 + i) − 1] .

1500
P
No ie that f ( ) = A − [(1 + )n − 1] = 750000 − (i/12) (1 + i/12)(12)(20) − 1 i .
n
ated.67%..The Seant
method would a likely . The minimal annual is approximatelyisompli6
f
26 Exerise Set 2.3
to a patient produes a onentration the blood stream given
28. A drug t/3
by
SOLUTION:0 baonen..25Assuming WhatWhenm g/mL? radministeredamountb etionshouldgiven?75%a.1shoulda nTheoft hemaximumb eoriginal.hoursinjet edmo untaft ero nent rationamouto rea hb eunitsisadmi nistered,t hishaveo urssafebeenlevel,whenifadminist eredinita ndt heis admi nisteredwhenseo nd.do esTheinjetion,thismaxim umwheno ur? whentlev elsafesho ulddropsat hirdto
−
c( ) = Ate A

injetion mg/mL, additional

0 = c′( ) = A 1 − 3 e− /3.

, we
need to administer =3 hours, and sine the onentration at this time will be c(3) = 3Ae −1

A= 1e

This happens whendetermine so that 3

1
This ours when 0.25 = c(t) = 3 te−t/3.
when
. We need to nd is 11 hours and 5 minutes; that is,t = 11.083 t hours.

b. We need to t so that units.

1
−t/ 3
29. ThisLet ours after21 hours and 0.5 = c( ) = 3 . + 0.75 3 e (t − 11.083)e−(t−11.083)/3.

14 minutes
3 +1 2
f (x) = 3 −7·5

ba. UsePlot the Maple omm nds.solve and fsolve to ry to nd all roots of f .
initialto nd the zeros ofto ro ts of .
. Newton's ndmethod f(x ) approximations f

.
d. Find the solu ions of f to within 10−16
f (x) = 0

exat
:= x > 33x+1 7 52x

solve − − · := x → 3(3x+1) − 7 52x

(f (x) = 0, x)

ln (3/7)
− ln (27/25)

fsolve SOLUTION:(f(x)= 0., xFirs) dene the funtionalgebraiallyby.

(3x+1) (2x)

The pro dure solve gives the exatfsolvesolution,(3 and −fsolve75 fails=beause0,x) the negative
asymptote for the graph of x -axis is an
f (x).
Solutions of Equations of One Variable
b. Using the Maple ommand plot({f (x)}, x = 9.5..11.5) produes the following graph.
y
27
15
. Dene 3 X10

2 X1015

1 X1015
10.5
11 11.5 12 x
for i from 1 to 5 dousing
′(x)
p1 := evalf(p0 − f (p0)/f p(p0))
p := x− > (D)(f )(x)
Digits f p := x → 3 3(3x+1) ln(3) − 14 5(2x) ln(5)
:= 18; p0 := 11 Digits
:= 18
p0 := 11

err := abs(p1 − p0)

p0 := p1
Theod results are given in the following table.

i pi |pi − p −1|
1 11.0097380401552503 0.0097380401552503
2 11.0094389359662827 0.0002991041889676
3 11.0094386442684488 0.2916978339 10−6
4 11.0094386442681716 0.2772 10−12
d. We have 5 11.0094386442681716 0

. Taking the natural logarithm of both sides gives

33x+1 = 7 · 52x

Thus (3x + 1) ln 3 = ln 7 + 2x ln 5.

7
and 3x ln 3 − 2x ln 5 = ln 7 − ln 3, x(3 ln 3 − 2 ln 5) = ln 3 ,
ln 7/3 ln 7/3 ln 3/7
This agrees with part (a). x =
ln 27 ln 25 = ln 27/25 = − ln 27/25 .
−
Exerise28 Set 2.4, page 85 Exerise

Set 2.4

1. wherea.Use Newton's method to nd a solution aurate to within 10−5 for x2 − 2xe−x + e−2x = 0,
SOLUTION:0≤x≤Sine1.

the iteration formula is

(x ) =x − −
2x e
2 x + e− 2x and f ′(x) = 2x − 2e− x + 2xe− x− 2e− 2x ,

or
aSOLUTION:.RepeatExer se 1(a)
p13 = 0.567135
Exeriseaurate to within 10−5

Continuingimprovement,we havespeed

Sine using Newton Raphson method desribed (2 13)

f (p ) 2 2p −pn−1 + −2pn−1

−
−1 n−1 −1

= p −1 − −
f ′(p −1) = pn−1 2p −1 − 2e−p + 2pn−1 −1 −pn−1 − 2e− p n−

With 0 = 0.5

manner,
in this 1 = 0.5 − (0.01134878)/(−0.3422895) = 0.5331555.
3. auraythemodiovedr -1? . in Eq. . . Is there

f (x) =x2 − 2xe− + e−2 ,

x x

f ′(x) =2x − 2e− + 2xe− − 2e−2 ,

x x x
and

f ′′(x) =2 + 4e− − 2xe− + 4e−2 ,

x x x
the iteration formula is
6. tha.Show p3 = 0.567143 is aurate to within 10−5

With f (pn−1)f ′(p −1)

required to hav pn = 1/n o n verg es linearly to , and determine the number of
p =0

− 2 .
pn = pn−1 [f ′(p −1)] − f (p −1)f
′′
(pn−1)

and p0 = 0.5 , we have f (p ) = 0.011348781, f (p ) = −0.342289542, f (p ) = 5.291109744

0
′
0
′′
0

Co t uingpin this=0.manner,5 (0.01134878)(−0.342289542) = 0.5680137.

 2
1 − (−0.342289542) − (0.011348781)(5.291109744)
n in Exerisethatth 1sequene. , whih is onsiderably beterms

First note that

linear |pn
.
− p| < 5 × 10−2 , we need 1/n < 0.05 , whih implies that
SOLUTION: |pn − p| < 5 × 10−2

lim 1 = 0. Sine
n→∞ n

pn+1 p 1/(n + 1) n
the onvergene is .To
lim | have − | = lim = lim = 1,

→∞ |pn − p| n→∞ 1/n →∞ n + 1

n > 20 .
 29

Solutions of Equations of One Variable

8. Show that:
b The sequene

a pn = 10−2
n nverges quadratially to zero;
k>1 .
SOLUTION:a.Sine pn = 10−n k does not onverge to zero quadratially, regardless of the size of

pn+1 0 10−2n+1 10−2n+1 10−2n+1

adratially .
bthe.For any i s qu| − | = lim = lim n = lim n+1 = 1,

→∞ →∞
l m

pn 0 2 →∞ (10−2 )2 10−2·2 n→∞ 10−2

sequene | − | onvergent
,
k>1

diverges. So the sp +1
0 10−(n+1)k 10−(n+1)k k k
−
lim | |
= lim = lim k = lim 102n −(n+1)
n→∞ |pn − 0|2 n→∞ 10−n
k
2 →∞ 10−2n n→∞

doesnot quadratially for any

method nk

10. Show that the xe d-point quene pn = 10− onverge .

k>1

mf (x)

has g(x) = x − f ′(x)

SOLUTION:g (p)=0,Ifif p is a zero of multipliityoftipliity .

′
m

f has a zero of m at p, then a funtion q exists with

Sine m where lim q(x) = .0

f (x) = (x − p) q(x), x p
6

expression x = p gives
have f ′(x) = m(x p)
m−1
q(x) + (x p) q′(x),
m
− −
m
whih redues to g( ) = x − mf (x) m(x − p) q(x)
=x ,
f ′′′
f ′(x) −
m(x p)m−1q(x) + (x p)m ′(x)

− −
Differentiating this andg(xevaluating)=x at m(x − p)q(x) .
− ′
−
mq(x) + (x )q (x)

If g′( ) = 1 mq(p)[mq(p)] = 0.
2
− [mq(p)]
are loseisontinuouenoughto,theTheontinuouslutionrem2.9implies that this sequene produes onvergene one we

12. Suppose that p.

only if f has m derivatives. Show that f has a zero of multipliity m at p if and

= f ( −1)(p),
m m
0 = f (p) = f ′(p) = but f ( )(p) = .0
··· 6
30 Exerise Set 2.4

SOLUTION: If f has a zero of multipliity m at p, then f an be written as

Thus f (x) = (x m
− p) q(x),
for x 6=p,
where lim q(x)
x p
6= .0

f ′(x) = m(x p) −1q(x) + (x p) q′(x)

m m
and
− −
f (p) = 0.′
Also
ad
′′
(x) = m(m − 1)(x − p) −2q(x) + 2m(x − p) −1q′(x) + (x − p) q′′(x)
m m m
′′
In general, f(p)
=for0.

k ≤m ,
k
k dj (x − p)m
X

(k)
f (x) = j q(k−j)(x)
j=0 j dx
(m − 1) th Taylor of about polynomial f ex panded p:

k
Thus, for = j=0 j m(m − 1)· · ·(m − j + 1)(x − p)m −j q(k −j)(x).
k

0≤k≤ −1 , we have f (k)

()=0 , but
(m)
Conversely, suppose that f (p) = m! x p
lim q(x) 6 = .0
→
′
f (p) = f (p) = . . . = f (m−1)(p) = 0 and f (m)(p) = .0Consider the
6

f ( −1)(p)(x − p) −1
′ m m m m
f (x) =f (p) + f (p)(x p) + . . . + + f ( )(ξ(x))(x − p)
− (m 1)! m!

(m) −
mf (ξ(x))
where =(x − p) m! ,
14. ShowisthatzerotheofSeantmethod onverges. of order
p multipliitym

is ontinuous, let
ξ(x) is between x and p. Sine f (m)
m
Then f ( )(ξ(x))
q(x) = .
m!

m
SOLUTION: Let sufiently
f (x) = (x − ) q(x) and
lim q(x) =
n , n+1
|e nα
| ≈ λ|e | . Thus
f ( )(p)
= .0
, the golden ratio.
So
x p
m! 6

→
| ≈ λ−1 |en|1 .
α /α /α
|en| ≈ λ|en−1| and |en−1

,where
α α= 1 + √ 5 /2
n = pn − p. If
then for large values of lim |e +1| = λ > 0,
α
n→∞ |e |
Solutions of Equations of One Variable
The hypothesis that for some onstant C and sufiently large n we have
31
|p +1 − p| ≈ C|p − p| |pn−1 − , gives
|
Sine the powersof λ e
α Ce λ−1/αe 1/α, so e α Cλ−1/α−1 1+1/α.

| n| ≈ | n| | | | | ≈ | n|
| n| must agree, √
Exerise2.ThisApplynumber,Newton'sSet2the.5,method Ratioto90approximate,α=1+ina1numerousroot/αofand α = 1+
in 5 . and in art.

Golden appears situations 2 mathematis

page
2x x
Generate terms until
x
f (x) = 6 + 3(ln 2)2 − ln 8 4 − (ln 2)3 = 0.
| n+1 − p | < 0.0002, and onstrut the Aitken's 2
sequene {pˆn}

Aitken's 2 Applying
3. onvergeLet quadratiallysequene, beausewehave sufisenta aurayrootwi.h pˆ 6 = −0.18 338 7. Newton's method fails to
0 = 0 requires nding p16 = −0.182888 . For the
SOLUTION: Newton's method wih
there multiple .

SOLUTION:g(x)=cos(Withx− 1) and p = 2. Use Steffensen's method to nd p . 0

(0)
0
(1)

g(x) = cos(x − 1) and p (0)

0= 2 , we have
 and p1(0) = g p0(0) = cos(2 − 1) = cos 1 = 0.5403023

Thus p2(0) = g p1(0) = cos(0.5403023 − 1) = 0.8961867.

What ould applied g(x) using p (0)

=1 and p (0)
=3 to obtain p (1)
.
= 0.75

p(1) = p(0) p1(0) − p0(0) 2

− (0) (0) )
0 0 (0) (0
2 − 2p1 − 2p1 + p0
(0.5403023 2)2
5. Steffensen's method is =2 1.173573 = 0.826427.
−
=2 to a funtion
−

− 0.8961867 2(0.5403023) + 2

p(0) be?
SOLUTION: Steffensen's method uses the formula
1
0 2 0

p(0) = p(0) p1(0) − p0(0) 2 .

−
1 0 (0) (0) (0)

p2 − 2p1 + p0
32 Exerise Set 2.5

Substituting for , , and gives

where

0
p(0) p(0)
2 0
p(1)
approx imate
10−5 of x = 0.5(sin x + cos x) ,
2 2
− (0)
(0) − 1
1 p1(0) − 1
Thus 0.75 = 1 , that is, 0.25 = (0) .
3 − 2p1 + 1 4 − 2p1
(0) (0)
1 (0) (0) 2 (0) (0) 2 (0)

and 1 − 2 p1 = 1 − 2p1 + 1, so 0 = p1 − 1.5p1 ,

11. b. Use p Steffensen's or method . to
=1.5 p =0 the solution to within
1 1

SOLUTION:g(x)= With0.5(sin x + cos x).

g(x) = 0.5(sin x + cos x), we have

p0(0) = 0, p1(0) = g(0) = 0.5,

p2(0) = g(0.5) = 0.5(sin 0.5 + cos 0.5) = 0.678504051,
−

p(1) = p(0) p1(0) − p0(0) 2 = 0.777614774,

0 0 (0) (0) (0)

p2 − 2p1 + p0
 p1(1)
p 2
=g p0(1)
1 = 0.
= 0.707085363, 2

(1) =g p(1)−
704939584,

p(2) = p(1) p1(1) − p0(1) = 0.704872252,

0 0 (1) (1) (1)

p2 − 2p1 + p0
p1(2) =g p0(2) = 0.704815431,
p2 1 = 0. 2

(2) =g p(2) 704812197,

p(3) = p(2) = p1(2) − p0(2) = 0.704812002,

0 0 (2) (2) (2)
p
any 2 − 2p1 + p0
that onvergesto
p = su perlinearly to 0, but does not onvergeof order for
α

and p1(3) = p0(3) = 0.704812002,

{p n } on verg es to with order , apositive onstant exists with
p α> 1 λ

Sine p2(3) = p1(3) = 0.704812197.

(3) (3) (3) , we aept
all agree to within as an answer that is
( 3)

urate to, within,and

0 2
−5
10 = 0.70 4812 197

p p p

2 1
sequene.

14. a a
10−5
b. Show that the {p } with order onverges superlin arly to .
1
α> 1 p

nn

λ = lim |pn+1 − p| .
α
|pn − p| n→∞
Solutions of Equations of One Variable
33
Hene
+
1 +1
This
l that
−
p
= lim | p
− α | p·n+1| p
− p α−1 =λ
· 0 =0 and lim −
p
= 0.
n→∞ p p →∞ pn p | n→∞ pn p
− | −
| −

b. Theimplies {p } that onverges p .

1
sequene onverges pn =
n
n
supe rlinearly to zero beause
n n
lim 1/(n + 1)( +1) = lim n
n
(n + 1)( +1)
n
n→∞ 1/n n→∞
n
n 1
lim
=
n→∞ n+1 n+1

However, for lim 1 1 1

= = · 0 = 0.
n→∞ (1 + 1/n)n n + 1 e
α> 1 , we have

1/(n + 1)( +1) nαn

lim n α
=lm ( +1)
→∞ (1/n ) n→∞ (n + 1)

n
n n(α−1)
m n+1 n+1
=
n→∞

So he does not f order 1 m (α−1)n = = .

=
n→∞ (1 + 1/n) n l→∞ +1 e ·∞ ∞
17. Let sequene nverge α for any α > 1.
a ForP ( ) n th Taylor polynomi l for f (x) = x
expanded about x0 = 0.

b. Let xed x, show that p = P (x) satises the hypothe thisof Theorem 2.14. n

Does Aitk en' s

x=1
, and use Aitken's 2 method to generate se quenepˆ0, pˆ1, . . . , pˆ8 .
SOLUTION: . 2 method the in situation?

be theSine aelerate onvergene

n 1 xk
pn = Pn(x) =
k!
k =0

X
x ξ
wherehave p p = P (x) e = −e xn+1,
n −n − (n + 1)!

Further,
ξ is between and . Thus, 0 x pn − p 6= 0 , for all . n≥0
ξ
−e 1 xn+2 (ξ1 ξ)
p n+1 − p (n+2)!
e − x
where = ξ = ,
pn p −e xn +1 n+2

− (n+1)!
ξ1 is between 0 and 1. Thus
 λ= lim e(ξ −ξ)x
1
= 0 < 1.
n→∞ n+2
 34 Exerise Set 2.6
b. The sequene has the terms shown in the following tables.

n 0 1 2 3 4 5 6

pn 1 2 2.5 2.6 2.7083 2.716 2.71805

pˆn 3 2.75 2.72 2.71875 2.7183 2.7182870 2.7182823

5 10− need p 7 8 9 10

n 2.7182539 2.7182787 2.7182815 2.7182818

page

. Aitken's pˆn 2.7182818 2.7182818

within method gives quite an improvement for this problem. For example, pˆ is aurate to 6
7 Exerise2.b.Use Newto n' sSet2.6,.m etho dWe to 1 00approximat e,tohav ethistoa uraywithin.

10−5 , the real zeros of

4 3 2 Appl yingpolyomialNewto n'slo met hodrdegree,withand determ ne any omplex zeros.

ThenSOLUTION:reduethe P (x) = x − 2x − 12x + 16x − 40.

4. approximatelyb.UseM¨uller's0method.58356±to1.nd49419the real. p0 = 1
and omplex zerosgivofs the sufiently aurate approximation

7 = −3.548233. When p0 = 4, we nd another zero to be p5 = 4.38 1113. If we divide by P (x )

we nd that (x + 3.548233)(x − 4.381113) = x2 − 0.832880x − 15.54521,

lists approximation
2 2

The omplex rootsP(xof) ≈the xquadrati−0.832880onthex right−15.an54521be x − 1.16712x + 2.57315 .

found by the quadrati formula and are

dueThe toseondthefatini[0t,ialhat2].approximationsThisisonrmedarebyusedthetoomplexndtherootsreal rooth thataregeneratedisknown.to between 4 and 5,
4 3 2

approxiSOLUTION:ationThewasfollowingusedbeausetable the P(x ) = x

−initial2 − 12 + 16 and−40the. roots. The rst ini ial
minimum in ,
f (0) = −40 f (1) = −37, nd f (2) = −56 implies that there is a
f (4) = −40 and f (5) = 115.
Solutions of Equations of One Variable
Thesinethird initial approximations are used to nd the real root that is known to lie between −3 and −4,
35
f (−3) = −61 and f (−4) = 88.

p0 p1 p2 Approximated Roots Complex Conjugate Root

0 1 2 7 = 0.583560 − 1.494188i 0.583560 + 1.494188i

2 3 4 6 = 4.381113

−2 −3

−4 points p5 = −3.548233
and use his information to sketh (thex) = x4 −f2 3 − 5 2 + 12x − 5, zeros of.

SOLUTION: There are at most real graph f

together with the fat that and , . Thi ,

5. b. Find the zeros and ritial four of f f (0) < 0 f (1) > 0 (2) < 0
in intervals lim x→∞ f (x ) = ∞ and , implies that these zeroslie
lim →−∞ f (x) = ∞

approxima ons these ,intervals , giv ,rosandthe . Applyingzeros:Newton's method,and various initial
(−∞ 0) (0,1) (1 , 2) (2, ∞) for

nd
ritial we need the ze of
points, approximate , ,
0.5798 1.521 2.332 , and −2.432. To
f ′(x) = 4x3 − 6x2 − 10x + 12.
to ndxthe= other1isquitetwo easilyzeros:seen to be a zero of f ′(x), the ubi equation an be redued to a quadrati
and to .

Sine the quadrati formula2applied−1.5

gives 0 = f ′′(x) = 12x2 − 12x − 10

√
sketh of the graphof , we also have the points of inetion.

A
x = 0.5 ± 39/6
f is given below.

y
60
9. Find a solution, aurate to within 40

20
22
x
23 21 1 2 3
220

10−4 , to the problem

4 3 2

by using the various 600methodsx− in550thisx hapter .

+2 00x
− 20x − 1 = 0, for 0.1 ≤ x ≤ 1
36 Exerise Set 2.6
SOLUTION: , we haveproblem
method:hod: For and , we have .
ba. BisetionNewtn's p0 = 0.1 p1 = 1 p14 = 0.23233
NotieSeant For p0 = 0.55, we have p6 = 0.23235.
of False Position: Forand .

d. ethod p0 = 0.1 p1 = 1 8 = 0.23235

. Muller's¨ : For p0 = 0.1 and p1 = 1, we have p88 = .

.230 25

theshapemethodandtheof BofirumferenesetionghtFalseirular,Positionmethodylindfor.morethismust have,wewasvolhaveonsiderablymeof1000 lessm .effetive than both the

11. SeantA inthatmethe p0 = 0 p 1 = 0.25, and 2 = 1 p6 = .23235

and h = 1000/ πr2 materialπ(+0.25)2

top and b ttom must ave a radius 3. To form seals, the

0.25 m than the radius nd the material f r the side must be ongerSinethan the volume is givenofbythe an. Minimize the amount of material that is r quired.

0 SOLUTION: ml .25

we have onstrut for . The amount f material req uired for the t
≈
V = 1000 = πr2h,
p of the an is
similar a mount is needed for the botto m. To onstrut the side of the an, the
needed, is

(2πr + 0.25)h. The total amount of material is given by M (r)

Thus M (r) = 2π r + 0.25)2 + (2πr + 0.25)h = 2π(r + 0.25)2 + 2000/r + 250/πr2.

(Fibonai) approximation
Solving M ′ (r) = 4π(r + 0.25) − 2000/r2 − 500/(πr3).
materialMneeded (r)=to0 ′
givesthean: Evaluating at this value of gives the minimal
r 5.3638 58 . M () r

12. Leonardo of Pisa foundMthe(5.base363858)60 ≈ 573.649 m . 2

as a root of the equation 1 + 22 60 + 7 60

2
+ 42 60
3 + 33 60 +4
4 5
60 + 40 60
6
1 1 1 1 1 1

equivalent?of Fibonai's base 60approximation is

3 2
HowSOLUTION:aurate wasThe hisdeimalapproximatio x + 2x + 10x = 20.

Newton's M thod gives 1.3688081078532 , and

rret
mlepalyno mialrnthm atiiandthe 3.2
within 1.36880810782137 with
10 − 11 .
tolerane f 10−16. So Fibona 's answer was F ib onailropoetof andprob aubily

tehniq uethatOmarisk nownKhay y´for approx imating toa mex ist,This.isleastthethis mostinEurope,ro otaur atefr om bef orethe apprwriting sthesix teeof thetoenturyg re atanirr ationPersi an.