You are on page 1of 28

Defence University; College of Engineering

CHAPTER 5
POWER SERIES

Introduction

In chapter 4 we were discussed about infinite series with constant terms. For the purpose of
major importance in applications we need to have series whose terms contain variables, usually
called power series. This chapter contains three sections. Power series and its basic properties of
convergence, Derivatives and Integration of power series, and Taylor series representation of
functions.

Objectives At the end of this chapter you will be able to


 know about power series
 explain about radius & interval of convergence of power series.
 find derivative & integral of power series.
 understand about Taylor series & Maclaurine series.

5.1. Power Series

Introduction
In this section we are trying to define a power series and discussing about the
convergence and divergence of a power series, and also find the radius and interval of
convergence of the power series.

Objective At the end of this section you will be able to


 define a power series.
 identify a region on which a power series is defined.
 state basic properties of a power series.
 determine whether or not a power series converges.
 find the radius of convergence of a power series.
 find interval of convergence of a power series.

Applied Mathematics I Power Series 403


Defence University; College of Engineering

Definition 5.1.1 A power series in − is a series of the form


 c n ( x  a ) n  c 0  c1 ( x  a )  c 2 ( x  a ) 2    c n ( x  a ) n  
n0

where are constants.


Note In the power series  ( − )
n 0

1. If a  0 , then the power series is given by


 c n x n  c 0  c1 x  c 2 x 2    c n x n  
n0

2. The value a is called center of the power series.


3. The values cn are the n th term coefficients of the power series.

Example 5.1.1 Find the center a and the coefficients c0 , c1 , c 2 and c3 of the following
power series.

xn
a. n0 n!

x 2 n 1
b.  (  1) n
n0 ( 2 n  1 )!

( x  5 ) 2 n 1
c.  (  1) n
n0 ( n  1) 5 n
 n
 x 
 ( 2 n )!  
 2 
d. n0

Applied Mathematics I Power Series 404


Defence University; College of Engineering

Solution
a. The center of the power series is a  0 and its first five coefficients of the terms are

1 1 1 1 1 1 1
c0   1 , c1   1 , c2   , c3   where cn  
0! 1! 2! 2 3! 6 n!
b. The center of the power series is a  0 and its first five coefficients of the terms are
1 1 1 1 1 1 1 1
c0   1 , c1     , c2   , c3   where cn  (1) n 
1! 3! 6 5! 120 7! 5040 (2n  1) !

c. The center of the power series is a  5 and its first five coefficients of the terms are

1 1 1 1 1 1 1
c0  0
 1 , c1   1
  , c2  2
 , c3   3

1 5 25 10 3 5 75 4 5 500

1
where c n  (1) n 
(n  1)5 n

d. The center of the power series is a  0 and its first five coefficients of the terms are

0! 2! 4! 6! ( 2n ) !
c0   1 , c1  1  1 , c2  2  6 , c3  3  90, where cn  n
2 0
2 2 2 2 

Example 5.1.2
Find value(s) of about which the following series are convergent
 
( x  3) n
a.  n!xn b. 
n0 n 1 n

Solution
n
a n 1  ( n  1) ! x n 1
a. Let a n  n ! x . Then and

a n1 (n  1) ! x n 1
L  lim  lim  lim (n  1) x  x lim (n  1)   , if x  0
n a
n
n  n! x n n  n 

Therefore, by generalized ratio test the series diverges if L  1 , that is the series

n
diverges for all real number x except x  0 . Hence the power series  n ! x
n0
converges only at x  0. 

Applied Mathematics I Power Series 405


Defence University; College of Engineering

( x  3) n ( x  3) n1
b. Let a n  , n  0 . Then a n 1  and
n n 1

a n1 ( x  3) n 1 n n
L  lim  lim  n
 lim ( x  3)
n a
n
n  n 1 ( x  3) n   n 1
n
 x  3 lim  x  3 , if x  3
n n  1

Therefore, by generalized ratio test the series

 diverges if L  1 , that is the series diverges for all real number


x  3  1  x  4 and x  2 and
 converges if L  1 , that is the series converges for all real number x 3 1 2  x 4 .

At this moment, the generalized ratio test gives no information about the convergence or
divergence of the power series at the boundary values x  2 and x  4 . So that, it is
necessary to check its convergence at x  2 and x  4 :


(  1) n
If x  2 , then the series  n 1 n
is a convergent alternating series.


1
If x  4 , then the series  n 1 n
is a divergent Harmonic series.


( x  3) n
Hence the power series  converges for 2  x  4 . 
n 1 n

Example 5.1.3

Find value(s) of about which the Bessel function J 0 ( x ) of order 0 is converges, where


x 2n
J 0 ( x)   ( 1) n
2 2 n n! .
2
n 0

x 2n x 2 n 2
Let a n  ( 1) n , n  0 . Then a n 1  ( 1) n 1 and
2 2 n n ! 2 2 n  2 ( n  1) !
2 2

2 2 n n !
2
a n 1 n 1 x 2 n 2 1
L  lim  lim (1)   lim x2
n  a
n
n 2 2n  2
(n  1) ! (1) x
2 n 2n n   4n  1
2

Applied Mathematics I Power Series 406


Defence University; College of Engineering

1
 x 2 lim 2
 0 , if x  0
n  4n  1

Therefore, by generalized ratio test the series

converges if L  1 , that is the series converges for all real number .



x 2n
Hence the Bessel function J 0 ( x)   ( 1) n of order 0 is converge for the set of all real
2 2 n n !
2
n 0

numbers. 

Lemma 5.1.1
 
a. If  c n s n converges, then  c n x n converges absolutely for s  x .
n0 n0
 
b. If  c n s n diverges, then  c n x n diverges for s  x .
n0 n0


Theorem 5.1.1 Let  c n x n be a power series. Then exactly one of the following
n0
conditions holds:

i.  c n x n converges only for x  0
n0

ii.  c n x n converges for all x
n0
iii. There is a number R  0 such that

 cn x n
n0 converges only for x  R
and

 cn x n
n0 diverges only for x  R

Applied Mathematics I Power Series 407


Defence University; College of Engineering

Remark
1. The number R in the above Theorem is called the radius of Convergence of the power

n
series  c n x and
n0
i. R  0 if the series satisfy (i)
ii. R   if the series satisfy (ii).

Thus, every power series have radius of convergence R , which is either non-negative number or
.

2. The collection of all values of x for which  c n x n converges is called Interval of
n0
Convergence of the power series.

3. If R is the radius of convergence of the series  c n x n , then the interval of convergence
n0

is one of the following intervals;  R , R  ,  R , R  ,  R , R  or  R , R  .

4. The most useful method of our disposal for determining the interval of convergence of the
power series is the Generalized Ratio Test.

Example 5.1.3 Determine the radius and interval of convergence for the following power series.


 1 n. n ( x  3) n 
( x  6) n
a. 
n 0 4n
d.  n n
n 1

2n 
x 2n
b.  ( 4 x  8) n e. 
n 1 n n 1 ( 3)
n


n 
xn
c.  n !( 2 x  1) f. 
n 0
n 0 n!

Solution


 1 n. n ( x  3) n , let  1 n (n) ( x  3) n . Then  1 n 1(n  1) ( x  3) n1
a. From 
n 0 4n
an 
4n
an 1 
4 n1
and

L  lim
a n1
 lim
 1 n1(n  1) ( x  3) n1

4n
n a
n
n 4 n1  1 n (n) ( x  3) n

 (n  1) ( x  3)
 lim
n  4n

Applied Mathematics I Power Series 408


Defence University; College of Engineering

n 1 1
 lim x3  x3
n 4n 4

Since by generalized ratio test

 if L  1 , the series will converge,


 if L  1 , the series will diverge, and
 if L  1 , the test is inconclusive.

then the series converges if

1
x  3  1  x  3  4  7  x  1  (1)
4

and the series diverge if

1
x  3  1  x  3  4  7  x and x  1  ( 2)
4

Thus, the radius of convergence for this power series is R  4 .

Now the interval of convergence (if not all) by solving the inequality in (1) above is

7  x 1

And then it is necessary to determine whether the power series will converge or diverge at the
endpoints of this interval.

If x  7 , then the series is given as


 1 n. n (  4) n 
 1 n . n (  1) n (4) n 


n 1 4n

n 1 4n
 n
n 1

This series is divergent by the Divergence Test since lim n    0 .


n

If x  1 , then the series is given as


 1 n . n ( 4) n 
 1 n. n (  1) n ( 4) n 
  (  1) n . n

n 1 4n

n 1 4n n 1

This series is also divergent by the Divergence Test since lim(  1) n . n doesn’t exist.
n 

Applied Mathematics I Power Series 409


Defence University; College of Engineering

Hence the interval of convergence is

 7  x  1   7 , 1 

2n
2n 2 n 1
b. By using the series  n
( 4 x  8) , Let a n  n
( 4 x  8) . Then an 1  ( 4 x  8) n 1
n 0 n n n 1
and

an 1 2 n 1 n 2n
L  lim  lim ( 4 x  8)n 1  n n
 4 x  8 lim  2 4x  8
n  a n n  1 2 ( 4 x  8) n   n 1
n

Now, by using generalized ratio test the series converges if 2 4 x  8  1 .

1
Since 2 4 x  8  1  8 x  2  1  x  2  then the radius of convergence for this power
8,
1
series is R  .
8

And, for interval of convergence

1
2 4x  8  1  8 x  2  1  x  2 
8

1 1 15 17
  x2  x
8 8 8 8

and then check for convergence at the end points:

15
 If x  , then the series is given as
8

n n n
2 n 15

n

2 n  1 
2 n  1 
 1

n 1 n
(
2
 8)  
n 1 n 
  
2
 
n 1 n 2 n
 
n 1 n

This is a convergent alternating series.

17
 If x  , then the series is given as
8

n
2 n 17

n

2n 1 
2n 1 
1

n 1 n
(  8)      
2 n 1 n  2  n 1 n 2
n
 
n 1 n

Applied Mathematics I Power Series 410


Defence University; College of Engineering

This is divergent harmonic series.

15 17 15 17 
Thus the interval of convergence for this power series is given as x  ,  
8 8 8 8


c. By using the series  n !( 2 x  1) n
, let a n  n ! ( 2 x  1) n . Then a n 1  n  1!( 2 x  1) n 1 By
n 0
using generalized ratio test you have

(n  1)!( 2 x  1) n1
L  lim  lim (n  1) ( 2 x  1)  2 x  1 lim (n  1)    1
n  n !( 2 x  1) n n n 

1
At this point we need to be careful, provided x   . So, this power series will only converge if
2
1
x   . Thus
2

The radius of convergence is R  0 and

1  1 1
The interval of convergence is x     ,   
2  2 2

( x  6) n ( x  6) n
d. By using the series  nn
n 1
, let a n 
nn
. Then by using Generalized Root Test

( x  6) n x6
L  lim n an  lim n
n
 lim 0
n  n n n n


( x  6) n
Since L  0  1 , then the power series 
n 1 n n
will converge for every x .

Thus,

 the radius of convergence is R   and


 the interval of convergence is    x   i .e.;  


x 2n x 2n
e. By using the series 
n 1 ( 3)
n
, l et a n 
(3) n .Then by using Generalized Root Test
x 2n x2 x2
lim n an  lim n  lim  
n n (3) n n  3 3

Applied Mathematics I Power Series 411


Defence University; College of Engineering

Now the series convergence if

x2
L 1  1  x 2  3  x  3   3  x  3
3

Therefore the radius of convergence is R  3 .

For determining interval of convergence:

 If x   3 , then the power series is



 3  2n



3 n 
   1
n
n n
n 1 (3) n 1 (3) n 1

n
which is divergent because lim  1 doesn’t exist.
n 

 If x  3 , then the power series is



 3 2n



3 n 

  1
n
and which is divergent.
n n
n 1 (3) n 1 (3) n 1

Thus the interval of convergence is  3  x  3   3 , 3   



xn xn x n 1
f. By using the series 
n 0 n!
, let a n 
n!
. Then an 1 
(n  1)!
and

an 1 x n 1 n! x
lim  lim  n  lim  0 .
n  a n   ( n  1)! x n   n
n


xn
Thus by Generalized Ratio Test the series 
n 0 n!
converges for all x.

Where

 the radius convergence is = ∞


 the interval of convergence (−∞, ∞) 

Exercise 5.1
1. Find the radius and interval of convergence of the power series given below:

Applied Mathematics I Power Series 412


Defence University; College of Engineering


1 
n2 n
a.  xn b.  n x
n 0 n  4 n 0 2
 
1 n 1
c.  (1) n 1
x g.  n
x 2 n 1
n 1 n n  0 ( 4)

n 
1
n 2
xn h.  (1) n 1 n 3 x n
d. n 0 1 n 1 3 n
 n
10 n

n 1 i.  x
e.  10 n
( x  4) n n  0 n!
n 0 
1

n! j.  n ( x  5) n
f.  100 n
xn n 1 n 5
n 0
2. Find the radius of convergence of the following power series

1 3 5  (2n  1) n 
nn n
a.  ( 1)n x b.  x
n 1 3 6  9  3n n 1 n !

a 1
3. If lim n 1  k , where k  0 , show that the radius of convergence of a n x n is .
n a k
n n 1


1
4. If lim n an  k , where k  0 , prove that the radius of convergence of a n x n is .
n
n 1 k
 
5. If a n x n has radius of convergence R, prove that a n x 2 n has radius of convergence R .
n 1 n 1

Applied Mathematics I Power Series 413


Defence University; College of Engineering

5.2 Differentiation and Integration of the Power Series


Introduction
In this section we will discuss derivatives and integrals of power series, and their application in
finding the power series representation of different functions. Actually, a function f defined by
a power series has similar properties to those of a polynomial. In particular, it can be shown that
f has a derivative whose power series representation may be found by differentiating each term
of the given series. Similarly, integrals of f may be obtained by integrating each term of the
series. These facts are consequences of the next theorems, which are stated without proof.

Objectives At the end of this section you will be able to

 find the derivatives of a power series.


 find the radius of convergence of derivatives of a power series.
 determine interval of convergence of derivatives of a power series.
 apply derivatives of a power series in solving real life problems.
 find the integral of a power series.
 determine interval of convergence of integrals of a power series.
 apply integrals of a power series in solving real life problems.

Theorem 5.2.1 (Differentiation of Power Series)


Let c n xn be a power series with radius of convergence R  0 . Then
n 0

the power series nc n x n 1 has the same radius of convergence and
n 1

d  n

d 

 n  
c x  c x n
  n c n x n 1
dx  n  0
n
n0 dx n 1

where x  R .

Applied Mathematics I Power Series 414


Defence University; College of Engineering

d   xn   xn
Example 5.2.1 Show that  
dx  n 0 n !  n 0 n !


xn
Solution From Example 5.1.3 (f) the power series 
n 0 n!
converges for all x .Thus by

Theorem 5.2.1 (differentiation Theorem of a power series)

d   xn   d xn 
n x n 1  x n 1 
xn
    
dx  n  0 n !  n  0 dx n ! n 1 n ! n 1 ( n  1) ! n 0 n !

d   xn   xn
Thus   
dx  n 0 n !  n 0 n !

Remark

xn
1. Let f ( x)   , for all x . Then f ( x )  f ( x) , for all x . Since the only function,
n0 n!
which is derivative of itself is e x and f (0)  e 0  1 , then f ( x)  e x , for all x and

xn
ex   .
n0 n!

2. By using (1) we have a. e 1  



 1n
n 0 n!

( x)n  (1)n n 2

x 2n
x
b.. e    x c. ex  
n 0 n! n 0 n! n 0 n!
 
3. Form Theorem5.2.1, the power series  an x n and  na n x n 1 have the same radius of
n0 n 1
convergence but they do not necessarily have the same interval of convergence.

xn  
For instance and  x n 1 have radius of convergence R  1 but their interval of
n 1 n n 1

convergence are  1 , 1 and  1 , 1 respectively.

Applied Mathematics I Power Series 415


Defence University; College of Engineering


Theorem 5.2.2 Suppose a power series c n xn has radius of convergence R0 .
n 0

Let f ( x )   c n x n . T hen f has derivatives of all order on  R , R  and
n 0

f ( n ) ( 0)  n ! cn , for all n  0.

Consequently,
 (n)
f ( 0) n
f ( x)   x , for  R  x  R.
n 0 n!

 
Corollary 5.2.1 Let  cn x n and
n 0
b
n 0
n x n be convergent power series for  R  x  R . If

 

 cn x n   bn x n for  R  x  R
n 0 n 0

then c n  bn for all n  0 .

Theorem 5.2.3 (Integration of Power Series)



Let  c n x n be a power series with radius of convergence R  0 . Then
n 0

cn n 1
 n 1 x
n 0
has the same radius of convergence and

x
   
c n n 1  x
0   c n t n  dt   t    c n t n dt for all x  R .
n 0  n 0 n  1 n 1 0


n
Recall A series of the form  C. x
n m
, where C is a constant is called a Geometric Series.

n
A geometric series  C. x
n m
for ≠ 0 converges if x  1 and diverges x  1 .

 
C xm
If the geometric series  C. xn converges, then
n 0
 C. x n 
nm 1 x
, x 1

Applied Mathematics I Power Series 416


Defence University; College of Engineering

Example 5.2.2 Show that

(1) n n1  (1) n1 n



ln(1  x)   x  x
n 0 n  1 n 1 n

n
Solution For all x  1 , the geometric series x
n 0
convergent and


n 1
x
n 0

1 x
…. (1)

By replacing x by  x in equation (1), we have got



n 1
 (  x)
n 0

1 x
…. (2)

Now, by using equation (2), we have got


x x 
1 n n

(1) n n1
ln(1  x )   dt    (1) t dt   x where x 1
0
1 t 0 n 0 n 0 n  1


(1) n n 1 x2 x3 x4
Thus, ln(1  x )   x  x     
n 0 n  1 2 3 4

Note The power series expansion of ln(1  x) for x  1 is sometimes known as Mercater’s
Series after the Danish Mathematician Nicolas Mercator, about 1620 – 1687.


(1) n 2 n 1
Example 5.2.3 Show that arctan x   x where x 1
n  0 2n  1

Solution If x  1 , then x 2  1 . Therefore, substitute  x 2 in place of x in equation (1),


you have got

1
2
  (1) n x 2 n for x 1.
1 x n 0

x  x
1 n 2n

(1) n 2 n1
Since arctan x   2
dt    (  1) t dt  
n  0 2n  1
x for x  1 , then
0 1 t n 0 0


(1) n 2 n1 x3 x5 x7
arctan x   x  x     
n 0 2n  1 3 5 7

Applied Mathematics I Power Series 417


Defence University; College of Engineering

(1) n 2 n 1

Note The series arctan x   x is called Gregory Series, after the
n 0 2n  1

Scottish Mathematician James Gregory, about 1638 – 1675.

Exercise 5.2 Determine power series representations for the following functions and its
interval of convergence:

1 x2 1
a. f ( x)  2
at c  0 e. f ( x )  at c  0
2  x  x 1
b. f ( x)  ln( 2  x ) at c  0 x
f. f ( x)  at c  0
1 2  3x
c. f ( x )  2 at c  2
x x
g. f ( x )  at c  0
1 1 x2
d. f ( x)  at c  0
x  12

5.3 Taylor Series and its Application


Introduction
In this section we will show how functions are infinitely differentiable, generate power series,
called Taylor series. In many cases, these series can provide useful polynomial approximation of
the generating functions. As we know that within its interval of convergence the sum of a power
series is a continuous function with derivatives of all orders. But in this section we will answer
the following basic questions. If a function f (x) has the derivatives of all orders on an interval I ,
can it be expressed as a power series on I ? And, if it can be, what will be its coefficients?

Objectives At the end of this section you will be able to

 find the coefficients of Taylor series of a function at a point.


 express a function as Taylor series expansion at a point.
 write a function as Maclaurine series expansion.
 identify the time where application of Taylor series expansion is applied.
 apply Taylor series expansion of a function in solving real life problems.

Applied Mathematics I Power Series 418


Defence University; College of Engineering

5.3.1 Taylor Series Expansion of a Functions

Definition5.3.1 Suppose that f has derivatives of all orders at 0 . Then the Taylor series
of f is the power series
 (n)
f (0 ) n
f ( x)   x .
n0 n!

Note 1. If we once find a power series representation of a function that power series must
be the Taylor Series for the function.

2. If f is a polynomial, say

f ( x )  c 0  c1 x  c 2 x 2    c n x n ,

then f is a power series with c j  0 , for all j  n , and f is its own Taylor
series.

Definition 5.3.2 For a function f that is not necessarily a polynomial, the n th  order
Taylor Polynomial Pn and the n th  order Taylor reminder rn of f are
defined by

f / ( 0) f // (0) 2 f ( n ) ( 0) n
Pn ( x)  f (0)  x  x  x
1! 2! n!

and rn  f ( x )  Pn ( x)

The Taylor Polynomials of f are the partial sums of the Taylor series of f

where

 (n)
f ( 0) n
f ( x)   x iff lim rn  0
n 0 n! n

Applied Mathematics I Power Series 419


Defence University; College of Engineering

Remark If we need to show, for a given value of x the Taylor series of f converges to f (x ) ,
we need to show that

lim rn  0
n 

To accomplish this we are usually use the Lagrange form of the remainder given in
Taylor Theorem below:

Theorem 5.3.1 (Taylor’s Theorem)

Let n be a non-negative integer and suppose f ( n1) ( x ) exists for each x in an


open interval I containing 0 . For each x  0 in I, there is a number t x
strictly lying between 0 and x such that

f // (0) 2 f (3) (0) 3 f ( n ) ( 0) n f ( n 1) (t x ) n 1


f ( x )  f ( 0)  f / ( 0 ) x  x  x  x  x . (3)
2! 3! n! (n  1) !

f ( n 1) (t x ) n1
and rn ( x )  x . .. (4)
(n  1) !

Note 1. From Taylor Theorem equation (3) is called Taylor formula and equation (4) is
called Lagrange Reminder formula.

2. If n  0 , then the Taylor formula becomes

f ( x )  f ( 0)
f ( x)  f (0)  x f / (t x )  x f / (t x )  f ( x)  f ( 0)  f / (t x ) 
x0

which is equivalent to the formula in the Mean Value Theorem for Derivatives.

Applied Mathematics I Power Series 420


Defence University; College of Engineering

Example 5.3.1 Show that

x3 x5 x7 
(1) n
sin x  x       x 2 n 1
3! 5 ! 7 ! n 0 ( 2 n  1) !

Solution Let f ( x)  sin x . Then f ( 0)  sin 0  0

1
f ( x )  cos x  f / (0)  cos 0  1 f (5)
( x )  cos x  f ( 5) ( x)  1

f 2  ( x )   sin x  f // (0)  0 f (6 )
( x )   sin x  f ( 6)
( x)  0
3 
f ( x )   cos x  f /// (0)  1 f (7 )
( x)   cos x  f ( 7)
( x)  1
( 4) ( 4) (8 ) ( 8)
f ( x )  sin x  f ( x)  0 f ( x)  sin x  f ( x)  0
( 2 k 1)
 f ( 2k )
( x)  (1) k sin x and f ( x )  (1) k 1  cos x , for some k  1, 2, 3, 


f ( n ) ( 0) n
Thus f ( x)   x
n 0 n!

f (0) 0 f 1 (0) 1 f 2  (0) // f ( 3) ( x )


 x  x  x  
0! 1! 2! 3!

1 3 x5 x7 
(1) n 2n 1
=0 x 0 x 0 0   x
3! 5! 7! n  0 ( 2 n  1) !

( n 1)
Since f (t x )  1 for all n  1 and tx is strictly lying between 0 and x , then

( n 1) n 1
f
(t x ) n 1 x
0  rn ( x)  x 
(n  1) ! (n  1) !

n 1
f ( n 1) (t x ) n 1 x
 0  lim rn ( x )  lim x  lim 0
n n  ( n  1) ! n  ( n  1) !

 lim rn ( x )  0  lim rn ( x )  0 (by squeezing Theorem)


n  n 

(1) n

Hence, f ( x )  sin x   x 2 n 1 
n  0 ( 2 n  1) !

Applied Mathematics I Power Series 421


Defence University; College of Engineering


f ( n ) ( 0) n
Note The Taylor series f ( x)   x is frequently called a Maclaurine series,
n 0 n!
after the Scottish mathematician Colin Maclaurine, about 1698 – 1746.

Important Maclaurine Series Interval of Convergence

1.

x 2 n 1
sin x   (1) n (  ,  )
n0 (2 n  1)!

2.

x 2n
cos x   (1)n (  , )
n0 ( 2 n) !

3.

xn
ex   (  , )
n 0 n!

4.

x n 1
ln(1  x)   (1) n (1 , 1]
n0 n1

5. 1 x 2 n 1

n
tan x   (1) [1 , 1]
n 0 2n 1

6.

x 2 n 1
sinh x   (  ,  )
n  0 (2 n  1) !

7.

x2n
cosh x   (  , )
n  0 ( 2 n) !

Table 5.1

5.3.2 Taylor Series about an Arbitrary Point

If a function f has a Taylor series

 ( n)
f ( 0) n
f ( x)   x  (5) ,
n 0 n!

then f must have derivatives at 0 and then it must be defined on an interval containing 0 .

Applied Mathematics I Power Series 422


Defence University; College of Engineering

For instance, since f ( x)  ln x is not defined on an open interval containing 0 , so that


f ( x)  ln x does not have a Taylor series of the form in (5). But as Example 5.3.1 the Taylor
series for f ( x )  ln( x  1) is

(1) n n 1
f ( x )  ln( x  1)   x , for x  1  (6)
n 0 n  1

Now, if 0  x  2 , then  1  x  1  1 , so that from equation (6)

( 1) n
ln x  ln 1  ( x  1)   ( x  1) n 1 , for 0  x  2
n 1

which is a Taylor series in powers of ( x  1) rather than powers of x and its interval of
convergence is centered at 1 rather than 0 .

Theorem 5.3.2 (Taylor Theorem about an Arbitrary Point)

Let n be a non-negative integer f  n1 ( x) exists for each n in an open interval I containing a . For
each x  a in I, there is a number t x strictly between x and a such that

f // ( a ) f ( n) (a) f ( n ) (t x )
f ( x)  f ( a )  f / (a )( x  a )  ( x  a) 2    ( x  a) n  ( x  a ) n 1  (7)
2! n! (n  1)!
and

f ( n 1) (t x )
rn ( x )  ( x  a) n1  (8)
(n  1)!

Note In the above Theorem, equation (7) is called Taylor formula and equation (8) is called
the Lagrange remainder formula.

Example 5.3.2 Express the polynomial f ( x )  2 x 3  9 x 2  11x  1 as a polynomial in ( x  3) .

Solution Let f ( x)  2 x 3  9 x 2  11x  1 . Then f (3)  5

f / ( x)  6 x 2  18 x  11  f / (3)  11 f /// ( x )  12  f /// (3)  12

f // ( x )  12 x  18  f // (3)  18
( n) (n)
f ( x)  0  f (3)  0 , for all n  4

Applied Mathematics I Power Series 423


Defence University; College of Engineering

Therefore,
 (n)
f (3)
f ( x)   ( x  3) n
n 0 n!

18 12
 5  11( x  3)  ( x  3) 2  ( x  3) 3
2! 3!

 5  11( x  3)  9( x  3) 2  2( x  3) 3 


Example 5.3.3 Find the Taylor series of sin x about and show that it converges to
6
sin x, for all x .

   1
Solution Let f ( x )  sin x . Then f    sin 
6 6 2

1    3
f ( x )  cos x  f 1    cos 
6 6 2

2     1
f ( x)   sin x  f 2      sin  
6 6 2
   3
f (3) ( x)   cos x  f (3)     cos  
6 6 2
   1  
f ( 4 ) ( x)  sin x  f ( x )  f ( 4 )    sin   f  
6 6 2 6


x
f t x  ( n 1)
 3
Since 0  rn ( x )   (x  )   f ( n1) t x   1, for all and
(n  1)! 3 (n  1)!


x
3
lim  0 , then lim rn ( x )  0
n   ( n  1)! n 

Thus the Taylor series of f ( x)  sin x converges to sin x . Moreover

( 2n )   n 1
f ( x )  (1) n sin x  f ( 2 n)
   (1) , for all  0 and
6 2

( 2 n 1) ( 2 n 1)   3
f ( x)  (1) n cos x  f    (1)
n
, for all  0
6 2

Applied Mathematics I Power Series 424


Defence University; College of Engineering

Hence,

 2n 
 (x  )  ( x  ) 2 n 1  3
sin x   (1) n 6   (1) n 6
n 0 2  ( 2 n ) ! n 0 2  (2n  1)!

1   1  3  
  (1) n  ( x  ) 2n  ( x  ) 2 n 1  
2 n 0  (2n)! 6 (2n  1) ! 6 

5.3.3 Application of Taylor Series

Example 5.3.4 Determine a Taylor Series about x  0 for the following integral.

sin x
 x
dx


( 1) n x 2 n 1
Solution By using the Taylor Series sin x   about x  0 ,
n  0 ( 2 n  1) !

sin x 1 1  ( 1) n x 2 n 1  ( 1) n x 2 n


 sin x   
x x x n  0 ( 2 n  1) ! n  0 ( 2 n  1) !

Thus

sin x 
(1) n x 2n 
(1) n x 2 n 
(1) n x 2 n1
 x dx   dx   dx  
 n0 (2 n  1)! n0  (2 n  1)! n0 (2 n  1) (2 n  1)!  C
sin x 
(1) n x 2 n1
 dx   C 
x n  0 ( 2 n  1) ( 2 n  1) !

Note

sin x
 Obviously you can’t integrate by using the integration technique by
x
substitution and by parts. So it is advisable to come up with a series representation
for the integral.

 This idea of deriving a series representation for a function instead of trying to find
the function itself is used quite often in several fields of study. In fact, there are
some engineering fields where this is one of the main ideas used and without this
idea it would be very difficult to accomplish anything in those fields.

Applied Mathematics I Power Series 425


Defence University; College of Engineering

Example 5.3.5 Find the first three non-zero terms in the Taylor Series for f ( x )  e x  cos x
about x  0 .

xn x
 
( 1) n x 2 n
Solution By using the Taylor Series expansions of e   and cos x   ,
n 0 n ! n 0 ( 2 n )!
you will have

  xn    (1) n x 2n 
e x  cos x       
 n 0 n!   n 0 ( 2 n) ! 

Now, write down the first few terms of each series and stop at the x 4 term in each.

 x 2 x3 x 4  x2 x4 x6 
e x  cos x  1  x      1     
 2 6 24  2 4 24 
2 4 3 5 4 6
 x x   x x   x2 x x 
 1       x          
 2 4 2 4 2 4 48
   
Second ser . 1 Second ser .  x x2
Second ser . 
2

 1 1  1 1  1 1 1 
Therefore, e x cos x  1  x      x 2      x 3      x 4  
 2 2  2 6  24 4 24 

x3 x4
 1 x 
 
3 6
 x3 x5 x7   x4 x 6 x8 
              
612 144
 
24 48 576
x3 x4
Second ser .  Second ser . 
6 24

1
Example 5.3.6  
Approximate  sin x 2 dx to four decimal places.
0

Solution By using the power series representation


x 2 n 1
n x3 x5
sin x   (1)  x  
n 0 (2 n  1)! 3! 5!

Applied Mathematics I Power Series 426


Defence University; College of Engineering

2x6 x 10
2 x 14
 sin x  x      ( 9)
3! 5! 7!

By applying term by term integration on the right side of the Taylor Series in equation (9), you
will get

1 1
 x3 x7 x11 x15  1 1 1 1
  2
0 sin x dx   3  3!7  5! 11  7! 15    3  42  1320  75600  
x 0

 sin x dx  0.3103


2
Thus the sum of the first three terms yields 
0

Example 5.3.7 Evaluate the following integrals.

 x2
a. e dx as an infinite series.
1
 x2
b. e
0
dx correct to within an error of 0.001.

Solution

xn
a. By using the Maclaurin series ex   and substituting  x 2 in place of x , the
n0 n!
2
Maclaurin series for e  x is given as

2

x 2n
e  x   ( 1) n  (10)
n0 n!

Now by applying term by term integration on equation (10), you will get

2n 2n
 x2  n x  
n x

( 1) n 2n1

 e dx    (1)
n! 
 dx    (1)
n!
dx    2n 1 n ! x 
 n 0 n 0 n 0 

Thus

2 n
x ( 1) 2 n 1
 e dx   2 n1 n! x  C
n 0

Applied Mathematics I Power Series 427


Defence University; College of Engineering

and it converges for all real number x .

b. By using the integral in (a)

1 1
 x2  x3 x5 x7 x9  1 1 1 1
e dx  x       1     
0  31 5 2 ! 7  3! 9  4!  x0 3 10 42 216

1 1 1 1
1     0.7475
3 10 42 216

The Alternating Series Estimation Method shows that the error involved in this approximation is less
than

1 1
  0.001 
11 5! 1320

Exercise 5.3

1. Find Maclaurine series representations for the functions below and determine values of x on
which it converges.
a. f ( x)  e3 x 1 d. f ( x)  tan 1 (5 x 2 )
b. f ( x)  cos(2 x   ) e. f ( x)  sinh x  sin x
x
c. f ( x)  cos 2 ( )
2
2. Find the Taylor series representations of the functions below about the given value of a .
a. f ( x )  e 2 x , a  1 1
e. f ( x)  2 , a  2
b. f ( x)  cos x , a   x
c. f ( x)  ln( 2  x) , a  2 f. f ( x)  x ln x , a  1

d. f ( x)  sin x  cos x , a 
4
3. By using the Taylor series generated by e x , show that

 ( x  a ) 2 ( x  a) 3 
e x  e a 1  ( x  a)    
 2! 3! 

4. By using Question 3, find the Taylor series generated by e x at a  1 .


5. For the following functions, approximate f by a Taylor polynomial with degree at the
number a .
a. f ( x)  x , a  4 and n  2

Applied Mathematics I Power Series 428


Defence University; College of Engineering

2
b. f ( x )  x 3 , a  1 and n  3
2
c. f ( x )  e x , a  0 and n  3
d. f ( x )  x sin x , a  0 and n  4

5.4 Unit Summary



 A power series can be written as  a n ( x  c) n .
n 0

 A power series in (1) is absolutely convergent if x  c  R and it is divergent if x  c  R .

 The value R in (2) is called the radius of convergence and any one of the following is called
interval of convergence for the power series ( − , + ), [ − , + ] , ( − , +
] or [ − , + ).

n
 Suppose a power series c x
n0
n has a nonzero radius of convergence R and let the function

f be defined by


f ( x)   cn x n  c0  c1 x  c2 x 2  c3 x 3  
n 0

for every x in the interval of convergence.

 If − < < , then

d 
i. f ( x )   cn x n  c1  2c2 x  3c3 x 2    n cn x n 1  
dx n  0
x  x
1 1 1 1
 f (t ) dt    c t n
n
 c0 x  c1 x 2  c 2 x 3  c 3 x 4    c n x n 1  
0 n 0 0 2 3 4 n 1

 The Taylor series representation a function f that has derivatives of all order at a point
a is given as

f n (a)
 n! (x  a)n
n0

( n 1 )
f ( a)
provided that rn ( x )  ( n 1) !
( x  a) n  0 as n  

Applied Mathematics I Power Series 429


Defence University; College of Engineering

 The Maclaurine series representation a function that has derivatives of all order at a point
is given as


f n (0)
 n! xn
n0

5.5 Review Exercises

1. Find the radius and interval of convergence for the following power series

( x  4) n 
xn
a.  c.  n
n 1 n 3n n 1 n


(1)n 1.(3 x  1) n 
(n  1) x 2n 1
b. 
n 1 n2
d. 
n0 3n

2. Find the Maclaurine series expansion for the functions


1 5
a. f ( x)  c. f ( x)  cos x 2
1  2x

b. f ( x)  sin x x
d. f ( x)  e 2

3. Find a series representation for


a. e  x in powers of x  2 .

b. x in powers of x  4 .
4. Using power series approximate values of the following problems in three decimal places.
1
a. 3
3e
1
sin x
b.  f ( x) dx , where f ( x )  if x  0 and f (0)  0
0 x

Applied Mathematics I Power Series 430

You might also like