Unit Two: Infinite Series

2. INFINITE SERIES
2.0 INTRODUCTION

The sum of infinitely many numbers may be finite. This statement which seems paradox at
first plays a central role in Mathematics and has a variety of important applications in
science and Engineering but many functions arising in applications are difficult to deal with.
One consequence of the early and dramatic success that scientists enjoyed when using
calculus to explain natural phenomena was that there suddenly seemed to be no limits, on
how infinite processes might be exploited. There was still considerable mystery about
“infinite sums” and “division by infinitely small quantities” in the years after Newton and
Leibniz, but even mathematicians normally insistent on rigorous proof were inclined to
throw caution to the wind while things were working. The result was a century of
exceptional progress in understanding the physical universe. Many mathematicians used
infinite series to derive previously inaccessible results. For example Laplace used infinite
series to prove the stability of the solar system.
In this unit we shall define infinite series, study how to determine the convergence or
divergence of an infinite series, state and prove some of the convergence tests. Estimate the
value of convergent infinite series.

Objectives:
At the end of this unit the student will be able to:
• define infinite series
• calculate partial sum
• identify convergent and divergent infinite series
• list tests that help to identify convergences or divergence of series.
• state generalized convergence test
• estimate the sum of a convergent series using integral test.
• estimate the sum of a convergent alternating series.
• apply different tests to determine convergence or divergence.
• identify the difference between infinite series and infinite sequences.

43
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

2.1 Partial Sums

Given any sequence {an }∞n = k , we can form a new sequence {sn }∞n = k as follows:

S1 = a1
S2 = a1 + a2
S3 = a + a2 + a3
.
.
.
Sn = a1 + a2 + a3 + ⋯ + an
Then s1 ,s 2, s3, ⋯ ,s n are called the partial sums and Sn is called the nth partial sums. This new

sequence {S n } is called a sequence of partial sums.

Is there any difference between infinite series and sequences?

Definition 2.1 Let {an } be a sequence of real numbers. The expression given by

∞
∑ an = a1 + a2 + a3 + ... + an + ... is called an infinite series (simply series) and the real
n =1

number an is called the nth term of the series.

∞
Notation: We shall use the notation ∑ an for the infinite series ∑a
n =1
n when there is no

possibility of confusion.

∞ ∞
Example 2.1 Let {an }n =1 = {n}n =1 be given. Then find the sequence of partial sum.

S1 = a1 = 1
S2 = a1 + a2 = 1 + 2 = 3
S3 = a1 + a2 + a3 = 1 + 2 + 3 = 6

44
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

S4 = a1 + a2 + a3 + a4 = 1 + 2 + 3 + 4 = 10
⋮
n ( n + 1)
Sn = 1 + 2 + 3 + 4 + ⋯ + n =
2
∞
∞  n ( n + 1) 
Therefore, the sequence {S n } of partial sum is given by { }
S n n =1 = 
 2 n =1

{( −1)n}
∞ ∞
Example 2.2 Let {an }n =1 = be given. Then find the sequence of partial sum.
n =1

S1 = a1 = −1
S2 = a2 = −1 + 1 = 0
S3 = a1 + a2 + a3 = −1 + 1 − 1 = −1
S4 = a1 + a2 + a3 = −1 + 1 − 1 + 1 = 0
⋮
−1 if n is odd
Sn = 
0 if n is even
∞
∞
 ( −1)n − 1
Hence, the sequence sn of partial sum is given by {sn }n =1 =  
2
 n =1
Remark: If Sn and Sn-1 are given the nth term a n = Sn – Sn-1.

Activity 2:1
∞
n −1
1. If the nth partial sum of a series ∑a
n =1
n is sn =
n +1
.
∞
Then find a. an b. ∑a
n =1
n

∞ ∞
n
2. If the nth partial sum of a series ∑ a n is sn = 3 −
n =1 2n
,then find ∑a
n =1
n

45
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

2.2 Convergence and Divergence

∞
Definition2.2 Let ∑ an be an infinite series with sequences of partial sums
n =1

S1, S2, S3 … Sn.

i) The series is said to be convergent if and only if lim s n is a real number
n→∞

in this case the number s = lim s n is called the sum of the series. We can
n→∞

∞
write s = ∑ an = a1 + a2 + a3 + ⋯ + an + ⋯
n =1

ii) The series is said to be divergent if and only if lim sn is not a real number
n→∞

or does not exist. A divergent series has no sum.

Example 2.3 Show that the series
∞ 1 1 1 1 1
∑n = + + +⋯+ + ...
n =1 ( n + 1) 1 ⋅ 2 2 ⋅ 3 3 ⋅ 4 n ( n + 1)
i) Convergent ii) find its sum.
Using partial fraction we have
1 A B
= +
n ( n + 1) n n +1

Multiplying both by n ( n + 1) we get

A ( n + 1) + Bn = 1

A+ B =0
Solving we find that A = 1 and B = −1 so

1 1 1
= −
n ( n + 1) n n +1

The sequence of partial sum will be

1
S1 = 1 −
2

46
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

1 1 1
S2 = 1 − + −
2 2 3
1 1 1 1 1
S3 = 1 − + − − =1−
2 2 3 4 4
⋮
1 1 1 1 1 1 1 1 1 n
Sn = 1 − + − + − + +⋯ + − =1− =
2 2 3 3 4 4 n n +1 n +1 n +1
∞
∞  n 
Hence, the sequence of partial sum is given by { }
S n n =1 = 
 n + 1n =1
Evaluating limit of partial sum we get

 
 n   1 
lim S n = lim   = lim  1  =1
n →∞ n →∞  n + 1  n →∞
1+ 
 n
∞ 1
It follows that lim S n = 1 . Thus the given series ∑ is convergent.
n →∞ n =1 n ( n + 1)
ii) The sum of convergent series is the limit of the partial sum
n
lim S n = lim =1
n →∞ n →∞ ( n + 1)

∞ 1
Therefore, S= ∑ =1
n =1 n ( n + 1)

Example2.4 Test the convergence of

∞ ∞ k2 − k −1
n −1
a. ∑ 2 b. ∑
n =1 k =1 k!
∞
Solution: Here a. ∑ 2n −1 = 1 + 2 + 22 + 23 + ... + 2n −1 + ...
n =1

The sums of the first few terms are

S1 = 1
S2 = 1 + 2 = 3
47
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

S3 = 1 + 2 + 4 = 7
⋮
sn = 1 + 2 + 22 + 23 + ⋯ + 2n −1
Then multiplying both side by 2 we get
2 3 4 n
2 S n = 2 + 2 + 2 + 2 + ... + 2
n
Sn = 2 S n − S n = 2 − 1

{ }
∞ ∞
Thus the sequence of partial sum Sn is given by {S n }n =1 = 2 − 1
n
n =1

and

(
lim S n = lim 2n − 1 = ∞
n →∞ n →∞
)
∞
Therefore the given series is ∑ 2n −1 is divergent.
n =1

b)
k 2 − k −1 k 2 k +1
= − =
k2
−
( k + 1) = k − ( k + 1)
k! k! k! k ( k − 1)! k! ( k − 1)! k !
The corresponding sequence of partial sum has the pattern
2
S1 = 1 − = −1
1!
2 2 3
S2 = 1 − + −
1. 1! 2!
2 2 3 3 4
S3 = 1 − + − + −
1! 3! 2! 2! 3!
⋮

Sk = 1 −
2
+
3
+
3
−
4
+
4
−
5
+ ... +
k
−
( k + 1)
1! 2! 2! 3! 3! 4! ( k − 1)! k!

Sk = 1 −
( k + 1)
k!

{ }
∞
∞ k +1
Thus the sequence of partial sum is given by S k
k =1
= 1− { }
k ! k =1
.

48
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

and

S = lim S k = lim  1 −
 ( k + 1)  = 1
k →∞ k →∞  
k! 
∞ k2 - k - 1
consequently the series ∑ is convergent.
k=1 k!
1
Definition 2.3 The series ∑n is called a telescoping series.
( n + 1)
Example 2.5 Find the sum of the series
3 5 7 9
+ 2 + 2 + 2 +⋯
1 1 + 2 1 + 2 + 3 1 + 2 + 32 + 42
2 2 2 2 2

Solution: Here the nth term of the series is

3 + 2(n − 1) 2n + 1 6 1 1 
an = = 6× = = 6 − 
1 + 2 + 3 +⋯ + n
2 2 2 2
n(n + 1)(2n + 1) n(n + 1)  n n +1
The sequence of partial sum is given by
 1 1 1 1 1 1 1 
sn = 6 1 −  + 6  −  + 6  −  + ⋯ + 6  − 
 2  2 3 3 4  n n +1 
 1 
= 6 1 − 
 n +1
∞
 6n 
Therefore the sequence of partial sum {sn }n =1 = 
∞
 and the sum of the series is equal to
 n + 1 n =1
6n
s = lim sn = lim =6
n →∞ n →∞ n +1
∞ 1
Example 2.6 Given the series ∑
n =1 ( 2n − 1)( 2n + 1)
a) find the sequence of partial sum
b) find its sum
1
Solution: Let an = . Using partial fraction we have
( 2n − 1)( 2n + 1)

49
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

1 A B
= +
( 2n − 1)( 2n + 1) 2n − 1 2n + 1

Multiplying both side by ( 2 n − 1)( 2 n + 1) and equating we find that

2 An + A + 2 Bn − B = 1

2 A + 2 B = 0
 ,
 A − B = 1
Solving A = ½ and B = -1/2
Therefore ,

1 1 1 1 
=  − 
( 2n − 1)( 2n + 1) 2  2n − 1 2n + 1 

Now we can calculate sums of few terms as follows

1 1
S1 = 1 − 
2 3

1 1 11 1
S2 = 1 −  +  − 
2 3 23 5

1 1 11 1 11 1
S3 = 1 −  +  −  +  − 
2 3 23 5 25 7
⋮
1 1 11 1 11 1 1 1 1 
sn = 1 −  +  −  +  −  + ⋯ +  − 
2 3 23 5 25 7 2  2n − 1 2 n + 1 
1 1 
= 1 − 
2  2n + 1 
∞
∞ ∞  n 
The sequence of partial sum { }sn n =1 is given by { }
S n n =1 = 
 2n + 1n =1
b) Since sum of convergent series is the limit of the partial sum
n 1
s = lim S n = lim =
n →∞ n →∞ 2 n + 1 2

50
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞ 1
The series ∑ is convergent and its sum S= ½
n =1 ( 2 n − 1)( 2n + 1)

 1
∞ ln 1 + 
Example 2.7 Given the series ∑  k
k = 2 ln k
ln ( k +1)
( )
a. Find the sequence of partial sum
b. Find its sum

 1  1  k +1
ln 1 +  ln 1 +  ln  
Solution: a. Let ak =
 k
=  k
=  k 
ln k ( ) (
ln k +1
)
ln ( k + 1) ln k ln ( k + 1) ln k

Simplifying using properties of logarithms we get

ln ( k + 1) − ln k ln ( k + 1) ln k 1 1
ak = = − = −
ln ( k + 1) ln k ln ( k + 1) ln k h ( k + 1) ln k ln k ln ( k + 1)

Listing some of the partial sum we see that

1 1
S1 = −
ln 2 ln 3
1 1 1 1
S2 = − + −
ln 2 ln 3 ln 3 ln 4
1 1 1 1 1 1
S3 = − + − + −
ln 2 ln 3 ln 3 ln 4 ln 4 ln 5
⋮
1 1 1 1 1 1 1 1
Sk = − + − + − + ... + −
ln 2 ln 3 ln 3 ln 4 ln 4 ln 5 ln k ln ( k + 1)

1 1
= −
ln 2 ln ( k + 1)

∞
∞  1 
The sequence of partial sum is given by S k = −
1
{ }
k =1  ln 2 ln ( k + 1)  
 k =1

51
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

 1 1  1
b. sum S = lim S k = lim  − =
k →∞ k →∞  ln 2 ln ( k + 1)  ln 2

 1
ln  1 +
∞  k 1
Thus we find that the sum of the series ∑ =
(
k = 2 ln k ln ( k + 1) )
ln 2

Exercise 2.1

1) For the following series

i. Find the sequence of partial sum
ii. Find its sum
∞  n +1 ∞ 1
a ) ∑ ln   c) ∑
n =1  n  n =1 n ( n + 2 )

∞ 4 ∞ 1
b) ∑ d) ∑
n = 0 16 n 2 − 8n − 3 n =1 n ( n + 1)( n + 2 )

2. Show that the following series are divergent

∞ 1 ∞ ∞ n ( n + 1)
b) ∑ ( −1)
k
a) ∑ c) ∑
n =1 n + n +1 k =1 k =1 2

52
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

2.3 Geometric Series

You are already familiar with the concept of geometric series from your high school
Mathematics. To recall what you have been learnt try the following activity.
Do you recall the definition of geometric series? When do we say that a geometric series is
convergent?

Activity 2:2
1. Suppose that you were to save one cent on September 1, two cents on September 2,
four cents on September 3, and so on, doubling the number each day. How much
would you save by the end of September?
2. A ball has bounce coefficient r < 1. If, when it is dropped from height h , bounce
back to a height of rh . Suppose such a ball is dropped from the initial height h and
subsequently bounces infinitely many times. Then Prove that the total up and
down distance it travels in all its bouncing is
 1+ r 
D =h 
 1− r 

∞
Definition 2.4 The series ∑ ar n −1 = a1 + ar + ar 2 + ... + ar n −1 + ...
n =1

is known as geometric series with common ratio r where a ≠ 0 .

Now we are at the position to stat and prove the following theorem.

∞
Theorem 2.1: The geometric series ∑ ar n −1 where a ≠ 0 is
n =1
∞ a
n −1
convergent if r < 1 and its sum is ∑ ar =
n =1 1− r
divergent if r ≥ 1 .

Proof: The nth partial sum of a geometric series is

S n = a + ar + ar 2 + ... + ar n −1
53
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

rS n = ar + ar 2 + ar 3 + ... + ar n
Multiplying both sides by r and subtracting we get

sn − rsn = a − ar n
sn (1 − r ) = a − ar n
a − ar n
Sn = ,r ≠1
1− r
a ar n
Sn = −
1− r 1− rn
To see convergence or divergence of the sequence of partial sums, we evaluate the following
limits by considering different cases on r.

 a ar n 
lim S n = lim  − , r ≠ 1
n →∞ n →∞ 1 − r 1 − r
 
a a
= − lim r n
1 − r 1 − r n →∞
The convergence or divergence of geometric series depends on the value of r.

n a
Case -1: If r < 1, then lim r = 0 . It follows that lim sn =
n →∞ n →∞ 1− r
∞ n −1 ∞ n −1 a
Thus the series ∑ ar is convergent and ∑ ar =
n =1 n =1 1− r

Case -2: If r > 1 , then

∞
n −1
a. for r > 1, we have lim r = ∞ and Implies the series ∑ ar
n
is divergent..
n →∞ n =1
n
b. for r<-1, then lim r = −∞ and lim sn = −∞
x →∞ n →∞

∞
n −1
Hence, the geometric series ∑ ar diverges.
n =1

Form a and b we can conclude that the geometric series ∑ ar n −1 is divergent if r > 1 .

Case -3: For r = 1 or − 1

54
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

a. If r = 1, then S n = a + a + a + ⋯ + a = an and lim sn = a lim n = ∞

n →∞ n →∞

0 if n is even
b. If r = -1, then Sn = 
− a if n is odd
Thus lim Sn does not exist. Hence,
n →∞
∑ ar n −1 is divergent for r = 1 or - 1.

Example2.8 Determine whether each of the following series converges or diverges. If it is

convergence find the sum
∞ k +2 ∞ n ∞
3 5 6
a. ∑   b. ∑ 4   c. ∑ n −1
k =1  5  n =1  2  n =1 4

Solution:
∞ k +2 3 4 5 k +2
 3  3 3 3 3
a. ∑   =   +   +   + ... +   + ...
k =1 5  5 5 5 5
3
3 3
is a geometric series with common ratio r = and a1 =   .
5 5
∞ 3 k +2
3  
Since r = < 1 , the series ∑   is convergent and its sum equal to
5 k =1 5 

3 3
3 3
∞ 3 k + 2    
   5  5 27
∑  5  = 3 = 2 = 50
k =1  1−
5 5
∞ n
5 5 52 53 5n
b) ∑ 4   = 4 ⋅ + 4 ⋅ 2 + 4 ⋅ 3 + ⋯ + 4 ⋅ n + ⋯
n =1  2  2 2 2 2
 5 52 53 5n −1 
= 10 1 + + + +⋯ + + ⋯
 2 22 23 2n −1 
 
∞ n −1
5
= ∑ 10  2 
n =1

55
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

5 5
is a geometric series with a = 10, and common ratio r = . Since r = > 1 , the series
2 2
∞ n
5
∑ 4  2  is divergent.
n =1  

∞ ∞ n −1
6 1 1
c. ∑ n −1
= ∑ 6  is a geometric series with a = 6 and common ratio r =
n =1 4 n =1  4  4
1
The given series is convergent since r = <1.
4
a
Example 2.9 Write the following decimal in the form of where a and b are relatively
b
prime.
a. 0.3ɺ = 0 ⋅ 3333333⋯

b. 2.3 45 =2.34545….
Solution: a) 0.3ɺ = 0.33333⋯
= 0 ⋅ 3 + 0 ⋅ 03 + 0 ⋅ 003 + 0 ⋅ 0003 + ⋯
 1 1 1 1 
= 3 + + + + ⋯
 10 100 1000 10000 
∞
 1 
= ∑ 3 n 
n =1  10 

1
is a geometric series with common ratio r = and a = 3 .
10
1 3 1 3 10 1
0.3ɺ = 0.3 = × = × =
1 10 9 10 9 3
1−
10 10

ɺ= 1
Therefore, 0.3
3
b) 2.345 = 2.3454545...
=2.3 + 0.045 +0.00045+ ⋯
= 2.3 + 0.045 + 0.00045 + 0.0000045 + ⋯

56
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

 1 1 1 
= 2.3 + 45  + + + ⋯
 1000 100000 1000000 
∞ n
 1 
= 2.3 + ∑ 45  
n =1  100 
1
Then it is a geometric series with common ratio r = and a=45.
100
1
23
2.345 = + 45 1000
10 1
1−
100
129
=
55
Example2.10 A young rabbit gives the tortoise a 100m head start. If a young rabbit runs at
1
5m/s and the tortoise at m/s how far the tortoise has traveled by the time a young rabbit
2
catches him?
Young rabbit

Let d1 be the distance tortoise traveled while a young rabbit was running the 100m to the
tortoise’s starting point.
 100m  1 
d1 =   m / s  = 10m .
 5m / s  2 
Let d 2 be the distance tortoise traveled while a young rabbit was running the distance d1 .

Since d1 =10 and a young rabbit runs at 5m/s,

 10m  1 
d2 =   m / s  = 1m .
 5m / s  2 

57
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

 d  1  d
In general, for each n, d n =  n −1    = n −1 .
 5   2  10
The total distance traveled by the tortoise is
∞
1
d = ∑ d n = 10 + 1 + +⋯
n =1 10
1
Which is a geometric series with a = 10, r = . This converges to
10
10 100
d= = m
1 9
1−
10
Example 2.11 Find the sum of infinite series
2 3
 1  1  1
1 + 2 1 −  + 3 1 −  + 4 1 −  + ⋯
 n  n  n
1
Solution: Let x = 1 − such that x < 1 and s be the sum of the infinite series
n
s = 1 + 2 x + 3 x 2 + 4 x3 +⋯
xs = x + 2 x 2 + 3 x 3 + 4 x 4 +⋯
(1 − x ) s = 1 + x + x 2 + x3 +⋯
This is a geometric series with common ration r = x < 1 . By theorem 2.1 the given series

converges to the sum

1 1
s= = = n2
(1 − x )
2 2
  1 
1 − 1 − n  
  

58
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞ ∞
Theorem 2.2: If ∑ an and ∑ bn are convergent infinite series with sums s and t
n =1 n =1
respectively and α ∈ ℝ \ {0} , then

∞ ∞ ∞
i. ∑( )
an + bn = ∑ an + ∑ bn converges to s + t
n =1 n =1 n =1
∞
ii. ∑ α an Converges to αs
n =1

∞ ∞
Proof: i. Let {sn } and {tn } be the sequence of partial sum for ∑ an and ∑ bn
n =1 n =1

∞ ∞
respectively. From hypothesis we are given that ∑ an = s and ∑ bn = t also from
n =1 n =1

sequence of partial sum we know that lim sn = s and lim tn = t . Further more {sn + tn } is the
n →∞ n →∞

∞
sequence of partial sum for the infinite series ∑ an + b
n =1
( n ) and

lim ( sn + tn ) = lim sn + lim tn = s + t .

n →∞ n →∞ n →∞

∞ ∞ ∞
Thus the series ∑( )
an + bn = ∑ an + ∑ bn converges to s + t .
n =1 n =1 n =1

∞
ii. Let {α sn } be the sequence of partial sum for the infinite series ∑ α an . Then
n =1

∞
lim α sn = α lim sn = α s . Hence
n →∞ n →∞
∑ α an converges to α s .
n =1

59
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

∞ ∞
Theorem 2.3 Let ∑ an be convergent series and ∑ bn be divergent series
n =1 n =1
If α ∈ ℝ \ {0} is any constant, then
∞
i. ∑ ( an + bn ) is divergent series.
n =1
∞
ii. ∑ α bn is divergent series.
n =1

∞
Proof: i. Suppose that the infinite series ∑ ( an + bn ) is convergent.
n =1
Let bn = (an + bn ) − an . Then
∞ ∞ ∞

∑ b = ∑ (a
n =1
n
n =1
n + bn ) − ∑ an .
n =1
∞ ∞
By hypothesis the series ∑ an is convergent series and assume that ∑ ( an + bn ) is
n =1 n =1
∞
convergent. It follows that the infinite series ∑ bn is convergent. But this contradicts the
n =1
∞
hypothesis. Thus ∑ ( an + bn ) is divergent series.
n =1
∞ ∞
ii. Let {sn } be a sequence of partial sums for the infinite series ∑ bn . Since ∑ bn is
n =1 n =1
∞
divergent series lim sn does not exists. Now assume that
n →∞
∑ α bn is convergent series. Then
n =1
1 1
lim α sn exists. But sn = (α sn ) , α ≠ 0 . Therefore, lim sn = lim (α sn ) = nlim sn exists.
n →∞ α n →∞ n →∞ α →∞

∞
Contradiction! Hence, ∑ α bn is divergent.
n =1
Note1. Changing the index of an infinite series does not affect it’s convergent or
divergence. But the sum in the case of convergent will be affected

60
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞ ∞
2. If ∑ an and ∑ bn are two series differing in their first few terms, then either
n =1 n =1

both convergent or both divergent

Harmonic series

∞
1
Theorem 2.4: The harmonic series ∑n
n =1
is divergent.

Proof: Let {sn } be the sequence of partial sum. Then

1 1 1
sn = 1 + + +⋯ +
2 3 n
1 1 1 1 1 1
s2 n = 1 + + + ⋯ + + + +⋯
2 3
 n n +1 n + 2 2n
sn

Subtracting we get
1 1 1 1 1 1 1 1
s2 n − sn ≥ + +⋯ ≥ + + +⋯ + =
n +1 n + 2 2 n 2n 2 n 2 n 2n 2
Implies
1
s2 n ≥ sn + ∀n > 1.
2
In particular
1 3 1
s22 ≥ s2 + ≥ + =2
2 2 2
1 1 5
s23 ≥ s4 + ≥ 2 + =
2 2 2
1 5 1
s24 ≥ s8 + ≥ + = 3
2 2 2
⋮
p+2
s2 p ≥ ∀p > 1
2
Hence, the sequence {sn } of partial sums can not have an upper bound and hence
∞
1
{sn } is divergent sequence. Consequently the series ∑ is divergent.
n n =1

61
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Theorem 2.5(Necessary Condition for Convergence of a series)

∞
If ∑a
n =1
n converges, then lim an = 0
n →∞

Proof: Let the n th and ( n -1) partial sums be as follows.

th

S n = a1 + a2 + a3 + ⋯ + an

S n −1 = a1 + a2 + a3 + ⋯ + an −1
Taking difference we find
S n − S n −1 = an
∞
From convergence of ∑a
n =1
n it follows that

lim S n = s
n →∞

Also from uniqueness of limit we get

lim an = lim ( S n − S n−1 ) = lim S n − lim S n−1 = s − s = 0 .
n →∞ n →∞ n →∞ n →∞

Remark: The converse of the above theorem is not necessarily true that means
lim an = 0 does not implies ∑ an converges.
n →∞

∞ 1
Example 2.12 a) Consider the Harmonic series ∑n
n =1

1 ∞ 1
lim an = lim
n →∞ n →∞ n
= 0 . But ∑n is divergent
n =1

∞ 1 1 1 1 3 1
b) ∑ n −1
= 1 + + + ... =
3 9
= convergent and lim an = lim n −1 = 0
1 2 n →∞ n →∞ 3
n =1 3 1−
3

62
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

From this example we observe fact that lim an = 0 does not indicate any thing about the
n →∞

∞
convergence or divergence of ∑ an .
n =1

Theorem 5.6: (The nth term Test for Divergence)

∞
If the limit lim an ≠ 0 or does not exist, then the series
n →∞
∑ an diverges.
n =1

Proof: (Contra positive of the above theorem)

Example 5.13 The following series are divergent
∞
a. ∑ 3 , sin ce ak = 3 and lim ak = lim 3 = 3 ≠ 0 .
k =1 k →∞ k →∞

Therefore, by nth term divergence test, it is divergent.

∞ 2k + 1 2k + 1
b. ∑ , sin ce ak = → 2 ≠ 0 . It is divergent by nth term divergence test.
k =1 k k

∞ n n +1
c. ∑ ( −1)
n =1 2n + 3
n
Here an =
( −1) n + 1
. and lim | an |= lim
n +1 1
= ≠0 .
2n + 3 n →∞ n →∞ 2 n + 3 2
∞ n +1
∑ ( −1)
th n
By the n term test for divergence is divergent.
n =1 2n + 3
Remark ∑ ( an ± bn ) converges does not implies ∑ an and ∑ bn converges.

∞ ∞ ∞
1 1 1 1 
Example 5.14 ∑ and ∑ n + 1 are divergent but ∑  n − n + 1  converges to 1
n =1 n n =1 n =1 

63
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Exercise 2.2
1. Prove that the following series are divergent

∞ n 2 + 2n ∞ ∞
1
∑ ( −1)
n
a) ∑ 2 b) ∑ n sin  n  c)
n =1 n + 5 n =1   n =1

2. The figure below show the first five terms of a sequence of squares. The outer most square
has an area of K m2. Each of the other squares is obtained by joining the mid points of the
sides of the squares before it. Find the sum of the areas of all the squares.

a
3. Show that the formula , when applied to the series 2 + 4 + 8 + 16 +⋯ , gives a negative
1− r
answer. Explain this seeming contradiction.

4. Find the sum of the series

1 1 1 1 1 1 1
1+ + + + + + + +⋯
2 3 4 6 8 9 12
Where the terms are the reciprocals of the positive integers whose only prime factors are 2s
and 3s.
∞
5. Give an example of a convergent series ∑a
n =1
n and bounded sequences {bn } such that

∑a b
n =1
n n diverges.

64
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞
6. Give an example of a series ∑a
n =1
n such that the series

∞
( a1 + a2 ) + ( a3 + a4 ) + ( a5 + a6 ) + ⋯ + ( an + an+1 ) + ⋯ = ∑ bn
n =1

∞
where bn = a2 n−1 + a2 n converges but ∑a
n =1
n diverges.

2.4 Nonnegative Series

In the previous section we have seen about convergence and divergence of a series by
actually computing its value. However, for most convergent series the exact sum is
impossible to find and it is suffice to know at least that the series converges. Under this
subsection we shall treat some common tests for convergence like Integral test, Comparison
test, Limit comparison Test, Ratio Test, and Root Test.
∞
Definition2:6 A series ∑a
n =1
n is said to be a nonnegative term series if an ≥ 0 for all n

element of natural number. If an > 0 for all n element of natural number, then ∑a n is called

a positive term series or a positive series.

∞
Let ∑a
n =1
n be a nonnegative term series. Then 0 < s1 < s2 < ⋯ sn ⋯ since sn +1 = sn + an+1 > sn .

That is the sequence of partial sums {sn } is strictly increasing. If the nonnegative term series
∞ ∞

∑ an has nth partial sums sn , then

n =1
∑a
n =1
n converges if and only if there exists K ∈ ℝ such

∞
that sn ≤ K ∀n ∈ ℕ. Further more if the nonnegative term series ∑a
n =1
n converges and has

sum s , then sn ≤ s ∀n ∈ ℕ or
k ∞

∑ a <∑ a
n =1
n
n =1
n ∀n ∈ ℕ.

65
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

5.4.1 The Integral Test

Theorem 2.7 Let {an }n =1 be a nonnegative sequence, and f be continuous,

∞

decreasing function defined on [1, ∞) such that f (n) = an for n ≥ 1 .

∞ ∞
Then the series ∑ an converges if and only if the integral ∫ f ( x)dx converges.
n =1 1

Proof: Consider the interval [1, n] and let P= {1, 2, 3, .n} be partition of [1, n]

Upper sum of f on [1, n] with respect to partition p is

f (2). (2-1) + f(2) (3-2)+…+ f(n-1) (n-(n-1)) = f(1) + f(2)+…+ f(n-1-1)

n −1 n −1
= ∑ f (i ) = ∑ ai = Sn −1
i =1 i =1

Lower sum of f on [1, n] with respect to partition p is

= f ( 2 )( 3 − 2 ) + f ( 3)( 4 − 3) + ... + f ( n ) ( n − ( n − 1) )

= f ( 2 ) + f ( 3) + ... + f ( n )

n n
= ∑ f ( i ) = ∑ ai = Sn − a1
i =2 i=2

66
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∴ S n − a1 ≤ ∫1 f ( x ) dx ≤ Sn ∀n = 1, 2,3...
n
(1)

Suppose ∑ ak converges say to S i.e. lim Sn = S

n →∞

From (1) taking limit as n → ∞ we obtain

lim ( sn − a1 ) = lim ∫ f ( x ) dx ≤ nlim ( Sn )

n
n →∞ n →∞ 1 →∞

Implies

∞
( s − a1 ) ≤ ∫1 f ( x ) dx ≤ s

∞ ∞
That is ∫1 f ( x ) dx is bounded. Thus, the improper integral ∫1 f ( x ) dx is convergent

(Because it is decreasing and bounded)

∞
Conversely suppose ∫1 f ( x )dx converges say to L . Then form (1)

∞
Sn − a1 ≤ ∫1 f ( x ) dx ≤ ∫1 f ( x ) dx = L
n

0 < Sn ≤ L + a ∀n = 1, 2....

∞
The sequence { Sn } is convergent (it is increasing and bounded). Thus the series ∑ an
n =1
∞ ∞
converges. Hence we can deduce that ∑ an converges if and only if ∫1 f ( x )dx converges
n =1

Example 2.15 Test the following series for convergence

∞ 1 ∞ ∞ 1 ∞ arc tan n
a) ∑ b) ∑ ne− n c) ∑ d)∑ 2
n =1 n n =1 n = 2 n ln n n =1 1 + n

Solution

67
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

1 ∞ 1
a) an = . The sequences {an }n =1 is nonnegativefor all n. Let f ( x ) = , ∀x ∈ [1, ∞ )
n x
. Then f ( n ) = an , for all n = 1, 2,....

−1
Trivially f is continuous on [1,∞ ) , f ' ( x ) = 2 < 0 ∀x ≥ 1 it follows that f is
x
decreasing ∀x ∈ [1, ∞ ) , hence by the integral test
∞ 1 ∞
∑ n = ∑ f ( n ) converges iff ∫1∞ f ( x )dx converges . But
n =1 n =1
∞ n1
∫1 f ( x )dx = nlim ∫
→∞ 1 x
dx = lim
n →∞
[ ln x ]n
1
= lim ln n = ∞ . This shows that the series
x →∞
∞ 1
∑
n =1 n

diverges

∞
b) an = ne − n .The sequences {an }n =1 is nonnegativefor all n.

Let f ( x ) = xe− x , ∀x ∈ [1, ∞ ) , then f ( n ) = a n , ∀n ∈ ℕ.

i) f is continuous on [1,∞ )

ii) ( )
f ' ( x ) = xe− x ' = e − x − xe− x = e− x (1 − x) ≤ 0 ∀x ≥ 1

Implies f is decreasing on [1,∞ ) then, by the integral test

∞ ∞ ∞
∑ ne− n = ∑ f ( n ) converges if and only if ∫1 f ( x )dx converges. But the improper
n =1 n =1
∞
∫1 f ( x )dx = tlim
t
integral ∫1
xe− x dx .Applying integration by parts we find that
→∞

−x
∫ xe dx = − xe− x + ∫ e− x dx = − xe− x − e− x + c

( )
t −x −x −x t
∫1 xe dx = − xe − e + c = −te − e + e + e
−t −t −1 −1

68
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

Taking limit yields

∞  2
∫1 xe− x dx = lim  te − t − e − t + 
t →∞  e
 t 1 2
= lim  −t , t + 
t →∞  e e e
t 1 2
= lim −t − lim t +
t →∞ e t →∞ e e

∞ 2
∫1 xe− x dx = <∞
e
∞
Thus the series ∑ ne− n is convergent by the integral test.
n =1

∞ 2
Can we conclude that ∑ ne− n = e ? Why?
n =1

1 ∞
c) an = . The sequences {an }n = 2 is nonnegativefor all n .
n ln n
1
Let f ( x ) = ∀x ∈ [ 2, ∞ ] Then f ( n ) = an , n = 2,3, 4....
x ln x

i) f is continuous on [ 2, ∞ )
− (1 + ln x )
ii ) f '( x) = < 0 ∀x ≥ 2 implies f is decreasing on [2, ∞ ).
( x ln x )2
All the hypothesis of the integral tests is satisfied, and then the
∞ 1
series ∑ converges iff ∫2∞ f ( x ) converges. Now we can evaluate the improper
n = 2 n ln n
integral by using integration by parts.

∞ 1 1
f ( x )dx = lim ∫2
t
∫2 dx . Let u = ln x, du = dx . Substituting we obtain,
t →∞ x ln x x

du = lim ( ln | u | |ln
ln 2 ) = lim ( ln(ln t ) − ln(ln 2) ) = ∞
∞ 1
∫2 f ( x ) dx = tlim
ln t t
→∞
∫ln 2 u t →∞ t →∞

69
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

The improper integral ∫2∞

1
dx is divergent. It follows that the infinite series
x ln x

∞ 1
∑
n = 2 n ln n

is divergent by the integral test.

arc tan n ∞
d) an = . The sequences { an }n =1
is nonnegativefor all n .
1 + n2

arc tan x
Let f ( x ) = ∀x ∈ [1, ∞ ) , then f ( n ) = an , n = 1, 2...
1 + x2

i) f is continuous on [1, ∞ )
1 − 2 x arc tan x
ii ) f '( x) = ≤ 0∀x ∈ [1, ∞ ) .
(
1+ x
2 2
)
Implies f is decreasing on [1, ∞ ) . Then

∞ arc tan x t arc tan x

∫2 2
dx = lim ∫1 dx
1+ x t →∞ 1 + x2

1
Now we apply integration by parts as follows. Let u = arc tan x, du = dx .
1 + x2
Substituting we get

arc tan t
∞ arc tan x arc tan t  u2 
∫1 dx = lim ∫π u du = lim  
1 + x2 t →∞
4
t →∞ 2
 π
4

2
π 
2  
 arc tan t   4 
= lim  −
t →∞  2  2

70
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

2 2
π  π 
    π2 π2 3 2
=  2
−  4
= − = π
2 2 8 32 32

∞ arc tan x 3π 2
∫ 2
dx = < ∞.
1
1+ x 32
∞ arc tan n
Implies the improper integral is convergent. Then the given series ∑ 2
is
n =1 1 + n
convergent by integral test.

arc tan n 3π 2 ∞
Is it possible to say the sum of the series ∑ 2
= ?
n =1 1 + n 32
∞
Let the positive term series ∑a
n =1
n be converges by the integral test. We wish to approximate

its sum by adding sufficiently many of its initial terms. The difference between the sum s and
the n th partial sum sn is the remainder

Rn = s − sn = an +1 + an + 2 + an +3 + … (1)

This remainder is the error made when the actual sum s is estimated by using in its place the
partial sum sn .

Theorem2: 8(Integral test Remainder Estimate)

∞ ∞
Suppose that the infinite series ∑a
n =1
n and the improper integral ∫
1
f ( x)dx

satisfy the hypothesis of the integral test, and suppose in addition that both
converge. Then
∞ ∞
∫
n +1
f ( x)dx ≤ Rn ≤ ∫ f ( x)dx
n

where Rn is the remainder given in (1).

71
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Proof: From adjacent figure we see that

k +1 k
∫k
f ( x)dx ≤ ak ≤ ∫
k −1
f ( x)dx

For k ≥ n + 1, n + 2, n + 3, ⋯ . We add these inequalities

for all such values of k and the result is the inequality

above because,

∞ k +1 ∞ ∞ ∞
∑∫ ∑∫
k
f ( x)dx = ∫ f ( x)dx and f ( x)dx = ∫ f ( x)dx .
k n +1 k −1 n
k = n +1 k = n +1

If we substitute Rn = s − sn , then the sum s of the series satisfies the inequality.

∞ ∞
∫ n +1
f ( x)dx ≤ s − sn ≤ ∫ f ( x)dx
n

Or
∞ ∞
sn + ∫
n +1
f ( x)dx ≤ s ≤ sn + ∫ f ( x)dx
n
( 2)
n +1
If the n th partial sum sn is known and the difference ∫n
f ( x)dx between the two integrals

is small, then (2) provides an accurate estimate of the actual sum s .

∞
1
Example2.16 Show that the series ∑ n(ln n)
n=2
2

a) Converges b) Find how many terms you would need to add to find its sum accurate to
0.01.

∞
. The sequences {an }n = 2 is nonnegativefor all n ∈ ℕ.
1
Solution: Here an = 2
n(ln n)
1
Let f ( x) = for all x ≥ 2 . Then f (n) = an , n = 2, 3, 4, …
x(ln x) 2

i) f is continuous on [2, ∞)

−(2 + ln x)
ii) f ′( x) = ≤ 0 for all x ≥ 2.
x 2 (ln x)3

72
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞
1
Thus f is decreasing on[2, ∞) . By integral test ∑ n(ln n)
n=2
2
Converges if and only if
∞
∫2
f ( x)dx converges. But

∞ t 1
∫2
f ( x)dx = lim ∫
t →∞ 2 x(ln x) 2
dx

Applying integration by substitution yields

ln t
∞ ln t 1  −1   − 1  1
∫ 2
f ( x)dx = lim ∫
t →∞ ln 2 u 2
du = lim   = lim 
t →∞
+  =
 u  ln1 t →∞  ln t ln 2  ln 2
< ∞.

∞ 1
Implies the improper integral ∫
2 x(ln x)2
dx is convergent. Thus by integral test

∞
1
∑ n(ln n)
n=2
2

converges.

b) To approximate the sum s , we need to choose n sufficiently large so that Rn < 0.01 i.e.

∞ 1 ln t 1 1
Rn ≤ ∫ 2
dx = lim ∫ 2
du =
n x(ln x) t →∞ ln n u ln n

Implies

1
Rn ≤ ≤ 0.01
ln n

Otherwise

ln n ≥ 100

Thus, n ≥ e100 ≈ 2.7 × 10 43 .

73
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

5.4.2 The P-series

∞ 1
Definition 5.7 The series ∑ p
where p > 0 is called p – series
n =1 n

1
Theorem2.9 Let p ∈ ℝ be fixed. Then ∑ Converges if p > 1 and
np
1
∑ diverges if 0<p ≤ 1.
np

1
Proof: Assume p > 0 and p ≠ 1 . Let f ( x ) = , ∀x ∈ [1, ∞ ) .Then
xp
i ) f Continuous on [1,∞ ) ii) f ( x ) > 0 ∀x ∈ [1, ∞ ) and
−p
iii) f ' ( x ) = p +1 < 0 ∀x ∈ [1, ∞ )
x
∞ ∞ 1
Hence by the integral test the series ∑ f ( n ) = ∑ p
n =1 n =1 n
∞
converges if and only if the improper integral ∫1 f ( x ) dx converges.
k
∞ 1  1   1  1− p
( x )dx
k
But ∫1 f = lim ∫ d x = lim  x 1 − p  = lim   k − 1 
k→∞ 1 xp k → ∞ 1 − p 1 k → ∞ 1 − p  

a) If 0 <p <1, then 1-p>0 and lim k

k →∞
1− p
= ∞ . Thus lim
1
k →∞ 1 − p
( )
k 1− p − 1 = ∞ and hence

1 ∞
the improper integral diverges. Hence, p ∑
diverges for 0 < p < 1 .
n =1 k
1− p  1 
b) If p > 1 , then 1 − p < 0 and so lim k = lim  p −1  = 0
k →∞ k →∞  k 

( )
∞
1 1− p 1
Therefore lim
k →∞ 1 − p
k −1 =
p −1
. Implies the improper integral ∫ f ( x)dx
1
converges.

∞ 1
Hence by the integral test the series, ∑ p
converges for p>1
n =1 n

74
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

1 ∞
Examples 2.17 a) The series 2 ∑
is convergent p-series with p =2>1
k =1 k
∞ 1 2
b) ∑ 2 is divergent by p-series test since p = < 1.
k =1 3
k 3
∞ 1 1
c) ∑ is divergent by p-series test since p = < 1.
2
k =1 k
Example 2.18 Give an example of two positive sequence {an } and {bn } such that
∞ ∞ ∞

∑
n =1
anbn < ∞ but both ∑ an and
n =1
∑b
n =1
n are divergent.

Solution: Take
1 1 if n is even
 if n is even 
an =  n 4 and bn =  1
1 if n is odd  n 4 if n is odd
∞ ∞
Both the series ∑ an and
n =1
∑b
n =1
n are divergent as lim an ≠ 0 and lim bn ≠ 0 by the n th term
n →∞ n →∞

1 1
divergence test. On the other hand anbn = 4
= 2 for all n ≥ 1 and
n n
∞ ∞
1
∑
n =1
anbn = ∑
n =1 n
2

converges.

75
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Exercise 2.3
1. Test the convergence of the following series
∞ 1 ∞ ln n ∞ 1
a) ∑ c) ∑ e) ∑
n =3 n ln [ 2n ]
2
k =1 k + 1 n =1 n
∞ 1 ∞ n
b) ∑ d)∑
k =1 k ln ( k + 1)
2
n =1 e n

2. Let { Pn } be an unbounded increasing positive term sequences.

∞
a) Prove that ∑( p
n =1
n +1 − Pn ) diverges.

∞
1 1 
b) Prove that ∑ P −  converges. Find its sum.
pn +1 
n =1  n

∞
1
3. For what values of p does the series ∑ n(ln n)
n=2
p
converges, and for what values of

p does the series diverges?

∞
1
4. Prove that the series ∑ (ln n)
n=2
p
diverges for all values of p .

76
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

2.4.3 Comparison Test

Theorem2:10 Let ∑a n and ∑b n be non negative term series such that

0 ≤ an ≤ bn ∀n ∈ ℕ.

∞ ∞
a) If ∑b
n =1
n converges, then ∑a
n =1
n converges.

∞ ∞
b) If ∑ an diverges, then
n =1
∑bn =1
n diverges.

n n
Proof: a) Let sn = ∑ ak and tn = ∑ bk be the sequence of partial sums of the series
k =1 k =1
∑a n

and ∑b n respectively.

Then {sn } and {tn } are increasing sequences since an ≥ 0 and bn ≥ 0 ∀n ∈ ℕ . Further more,
∞
0 ≤ sn ≤ tn since 0 ≤ an ≤ bn ∀n ∈ ℕ. Suppose that ∑b
n =1
n converges. Then
n
lim tn = lim ∑ bk exists. The sequence {tn } is increasing and bounded. It follows that the
n →∞ n →∞
k =1

sequences {sn } is also bounded and increasing. Hence, {sn } is convergent since every
∞
increasing and bounded sequence is convergent. Thus, the series ∑a
n =1
n converges.

∞
b) Suppose ∑a
n =1
n diverges. But from hypothesis 0 ≤ an ≤ bn ∀n ∈ ℕ. , then the sequences of

partial sums sn ≤ tn ∀n ∈ ℕ. Taking limit we get

lim sn ≤ lim tn
n →∞ n →∞

This implies
∞
∞ ≤ lim tn Since
n →∞
∑a
n =1
n diverges.

77
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

∞
Therefore, lim tn = ∞ . This show that the sequence of partial sums of the series
n →∞
∑b
n =1
n is not
∞
bounded. Hence, the series ∑b
n =1
n diverges.
∞
1
Example2.19 Prove that ∑ k ! is convergent.
k =1
∞
1
Proof: ∑ k ! is a series of nonnegative terms. We need to show that the sequence of partial
k =1

sums {sn } has an upper bound.

1 1 1
s1 = 1, s2 = 1 +
, s3 = 1 + + , ⋯,
1⋅ 2 1⋅ 2 1⋅ 2 ⋅ 3
1 1 1
sn = 1 + + +⋯ +
1 ⋅ 2 1⋅ 2 ⋅ 3 1 ⋅ 2 ⋅ 3⋯ n
∞
1 1
Consider a geometric series ∑ k −1 where a=1 and r = . So it is converges to 2 that is to
k =1 2 2
say

n 1 ∞ 1
∑ k −1
≤ ∑ k −1
=2
On the other hand
k =1 2 k =1 2
k ! ≥ 2k −1 ∀k = 1.2.3....
1 1 n 1 ∞ 1 ∞ 1
Taking reciprocal we get ≤ k −1 . It follows sn = ∑ ≤ ∑ k −1 ≤ ∑ k −1 = 2
k! 2 k =1 k ! k =1 2 n =1 2

The sequence of partial sums { sn } is convergent, since it is increasing and bounded

1
Hence, ∑ n! is convergent by comparison test
Example 2.20 Test the convergence of
∞ 1 ∞ k2
a) ∑ 3
c) ∑ 3
k =1 2k +1 k =1 k + 5k
∞ 1 ∞ 1
b) ∑ d) ∑
k =1 ( 3k + 1) k = 2 ln k
Solution:
a) k 3 < 2k 3 + 1 ∀k ∈ ℕ . Taking reciprocal we get
1 1
< 3
2k + 1 k
3

Implies
78
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞ ∞
1 1
∑ 2k
k =1
3
+1
<∑
k =1 k
3

1
But the series ∑
convergent p – series with p = 3>1.
k3
∞ 1
Therefore, by comparison test the series ∑ 3
is convergent
k =1 2 k + 1
b) For each k element of natural number, 3k + 1 < 3(k + 1) . Then
1 1
0< < ∀k ∈ ℕ
3 ( k + 1) 3k + 1
It follows that
1 ∞ 1 ∞
1
∑ < ∑
3 k =1 k + 1 k =1 3k + 1
∞
1
Since ∑ k + 1 is divergent series by integral test, it follows that by the comparison test the
k =1
∞
1
given series ∑ 3k + 1 divergent.
k =1

c) For each k element of natural number, k 3 + 5k < k 3 + 5k 3

Taking reciprocal and multiplying by k 2 we get
k2 1 k2
= <
k 3 + 5k 3 6 k k 3 + 5k
Taking summation we obtain
1 ∞ 1 ∞ k2
∑ <∑
6 k =1 k k =1 k 3 + 5k
∞ ∞
k2 1
∑
Therefore, the series
k =1 k + 5k
3
diverges by comparison test since ∑
k =1 k
is divergent.

1 1
d) ln k < k ∀k ∈ ℕ and < , ∀k = 2, 3, 4,⋯
k ln k
∞
1 ∞ 1
∑
k =2 k
<∑
k = 2 ln k
∞ ∞
1 1
Since ∑ is divergent, by comparison test ∑ is also divergent.
k =2 k k = 2 ln k
∞ ∞
Remarks: 1. If l ∈ ℝ , then ∑ an converges if and only if
n =1
∑ la
n =1
n converges.
∞ ∞
2. ∑a
n =1
n Converges if and only if ∑a
n=m
n converges.

79
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Theorem 2.11 (Limit comparison test)

Suppose that an > 0 and bn > 0 for all, n ∈ ℕ .
∞ ∞
an
i) If lim
n →∞ b
= c > 0, then ∑a
n =1
n and ∑b
n =1
n both converges or both diverges.
n

∞ ∞
an
ii) If lim
n →∞ b
= 0 and ∑ bn converges, then
n =1
∑a
n =1
n converges.
n

∞ ∞
an
iii) If lim
n →∞ b
= ∞ and ∑b
n =1
n diverges, then ∑a
n =1
n diverges
n

Example 2.21 Test the convergence of the following series

∞ 3k − 2 ∞ 1
1
a) ∑ 2
c) ∑ sin  
k =1 k + 1 n =1 n n
∞
b) ∑
n =1
( 4 4
n +1 − n −1 ) ∞ 1 + n ln n
d) ∑
n=2 n2 + 5
∞ n +1 ∞ 3 3
e) ∑ f ) ∑ 8k − 5k
n =1 3n + 2 n k =1
3k − 2 1
Solution: a) ak = 2
is nonnegative term for each k in ℕ and it behaved like for
k +1 k
1
large values of k. Now let bk = in order to compare with ak . Evaluating the following
k
limit we obtain
ak 3k − 2 3k 2 − 2k
lim = lim 2 ⋅ k = lim =3>0
k →∞ bk k →∞ k + 1 k →∞ k 2 + 1
∞ 1 ∞3k − 2
Since ∑ k diverges, then by limit comparison test ∑ 2
diverges
k =1 k =1 k + 1
b) an = n 4 + 1 − n 4 − 1 is nonnegative for each n ∈ ℕ and the given series is a series
of positive term.

n4 + 1 − n −1 =4
( n4 + 1 − n4 − 1 )( n4 + 1 + n4 − 1 )
n4 + 1 + n4 − 1

80
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

=
(
n4 + 1 − n4 − 1 ) =
2
=
2
n4 + 1 + n4 − 1 n4 + 1 + n4 − 1  1 1 
n2  1 + 4 + 1 − 4 
 n n 
1 1
For n → ∞, an be having like 2 . Choosing bn = 2 and compute
n n
 
 
= lim   ⋅ n2
an 2
lim
n →∞ bn n →∞   1 1 
 n2  1+ 4 + 1− 4  
 n n  
 
2
= lim
n →∞  1 1 
 1+ 4 + 1− 4 
 n n 
2
= =1
1+1
an
Thus, lim = 1 , 0 < 1 < ∞ . There for by limit comparison test, the series
n →∞ bn

∞ ∞
∑ an and ∑ bn converges or diverges together.
n =1 n =1

∞ 1 ∞
But the series ∑ 2 is convergent by p – series test. Hence, ∑ an is convergent
n =1 n n =1

1 1 1
c) an = sin   . is nonnegative term for each n. Take bn = 2 then we have
n n n
1 1 1
sin   sin
a n 2 n = 1 by L’Hopital’s rule which is finite non-zero
lim n = lim = lim
n →∞ bn n →∞ 1 n →∞ 1
n2 n
∞ ∞
1 1 1
number. Since ∑ 2 is convergent p-series, then the series ∑n sin   is convergent
n =1 n n =1 n
by limit comparison test.

81
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

1 + n ln n n ln n ln n 1 ln n
d) Let an = . For n large, an behaves like 2 = . But < for n ≥ 3 . So
n +5
2
n n n n
1 ∞ ∞
1 a n + n 2 ln
let bn =
n
. Since ∑ bn = ∑
n=2 n=2 n
diverges and lim n = lim 2
n →∞ b n →∞ n + 5
= ∞ , the series
n

∞ 1 + n ln n
∑ diverges by limit comparison test
n=2 n2 + 5
e and f are left as exercise.
∞
1
Example 2.22 Determine the convergence or divergence of the series ∑ .
n =1 1 + 2 + 3 + ⋯ + n
2 2 2

n(n + 1)(2n + 1) 2n3 + 3n 2 + n

Solution: We know that 1 + 2 2 + 32 + ⋯ + n 2 = = . Thus the nth
6 6
terms of the sequence reduced to
1 6
an = = 3
1+ 2 + 3 +⋯ + n
2 2 2
2n + 3n 2 + n
∞ ∞
1 1 1
Here for large n, an behaves like
n 3
. Let bn = 3 and
n
∑b = ∑ n
n =1
n
n =1
3
is convergent by p-series

test. Furthermore,
an 6n 3
lim = lim 3 =3> 0.
n →∞ b n →∞ 2 n + 3n 2 + n
n

∞
1
Hence by limit comparison test the series ∑ 1+ 2
n =1
2
+ 32 + ⋯ + n 2
converges.

∞ ∞
Theorem 2.12: Let ∑ an be a series of positive term and ∑ an be convergent.
n =1 n =1

∞
Then the series ∑ an2 is convergent.
n =1

∞
Proof: Given that ∑ an is convergent, then by divergent test lim an = 0
n →∞
n =1
By definition, for every ∈> 0 there is n ≥ N so that if n ≥ N , than | an − 0 |<∈
Let 0 <∈< 1 and consider | an − 0 | .Applying property of absolute value and simplifying we
get
82
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

−an < an2 < an

Implies
0 < an 2 < an why?
Taking summation we find
∞ ∞
∑ an2 < ∑ an
n =1 n =1
∞ ∞
Since ∑ an is convergent , by comparison test ∑ an2 is convergent.
n =1 n =1

Note: The converse of the above theorem is not necessarily true.

∞ 1 1 1
Example2.23. Let ∑n be given where an =
n
and an2 = 2
n =1 n
∞ ∞
Here the series ∑ an is divergent but ∑ an2 is convergent
n =1 n =1

∞
For a given series ∑ an , success using either comparison test to determine whether
n =1
∞ ∞

∑ an converges will depend on coming up with another series

n =1
∑b
n =1
n (geometric, p-series. …)

whose convergence is known and for which an and bn may be compared.

83
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

2.4.4 RATIO TEST

Applying the comparison test require several steps of trial and error to find a suitable
comparison. An alternative, the ratio test, enables a kind of internal comparison of terms of
the given series and has the added advantage of being applicable to series that contain
negative terms.

∞
Theorem2.13: Let ∑ an be a series of positive terms such that
n =1

a
lim n +1 = ℓ ( possibly ℓ = ∞ ) . Then the following hold
n →∞ an

i) if 0 ≤ ℓ < 1, then the series converges

ii) if ℓ > 1 , then the series diverges

iii) if ℓ = 1 then the test fails to indicate convergence or divergence

a
Proof: i) Suppose that lim n +1 = ℓ < 1 . Given ε > 0 there exists a number N > 0 such that
n →∞ an
an +1
− L < ε for all n ≥ N . That is
an
a
If n ≥ N , then l − ε < n +1 < l + ε . Now choose ε > 0 such that l + ε < 1 . Let r = l + ε , then
an
an +1
< r < 1 for all n ≥ N .
an
Thus,
a N +1
< r implies aN +1 < raN
aN

aN + 2
< r so a N + 2 < ra N +1 < r 2 aN
aN +1

84
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

aN +3
< r implies aN +3 < ra N + 2 < r 3a N
aN + 2
⋮
∞
This shows that for n ≥ N we can dominate ∑ an by a N (1 + r + r 2 + r 3 + ⋯) that is
n =1
∞ ∞

∑
n = N +1
an < ∑ r n a N
n =1
∞
Since ∑r a
n =1
n
N is a geometric series with common ratio r < 1 , it is convergent.

∞
Therefore, by comparison test ∑ an is convergent.
n =1
a a
ii) Suppose lim n +1 = ℓ > 1 . Then there exists N > 0 such that n +1 > 1 for all n ≥ N .
n →∞ an an
In particular,
aN < a N +1 < aN + 2 < ⋯
∞
The terms of the series do not approach 0, so ∑ an diverges by the n th term test.
n =1
∞ ∞
1 1
iii) Suppose l = 1 . The series ∑
n =1 n
is divergent and ∑n
n =1
2
is convergent. But in both cases

an +1
lim = 1.
n →∞ an
Therefore, the test fails to indicate the convergence or divergence.
What you will do if the ratio test fails to indicate convergence or divergence for a given
series?
Example 2.24 Test the convergence for the following series
∞ n! ∞
(3n)!+ 4n +1
a) ∑ b) ∑
n =1 (3n + 1)!
n
n =1 n
∞ ∞
4n n ! n ! 1 ⋅ 3 ⋅ 5 ⋅⋯ (2n − 1)
c) ∑ d) ∑
n =1 (2 n)! n =1 4 n 2n n !

Solution: Let an = n ,
n!
an +1 =
( n + 1) , then compute
n ( n + 1)n +1

85
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

an +1
=
( n + 1)! n n
. =
( n + 1) .n! n n  n n  1 
n

n +1
. =  = 1 −  .
an ( n + 1) n ! ( n + 1) ( n + 1)
n
n !  n + 1   n +1
x
 1 
Let f ( x) = 1 −  for all x ≥ 1 , then f (n) = an for all n ≥ 1 and
 x +1 
x
 1  ∞
lim f ( x) = lim  1 −  has indeterminate form of type 1 . Taking natural logarithm to
x →∞ x →∞
 x +1 
both side we obtain,
x
 1   1 
lim ln f ( x) = lim ln 1 −  = lim x ln 1 −  still has indeterminate form of type ∞ ⋅ 0
x →∞ x →∞
 x + 1  x →∞  x +1
 1 
ln  1 - 
lim lnf ( x ) = lim  x +1 0
has indeterminate form of type . Now applying
x →∞ x →∞ 1 0
x
L’Hopital’s Rule I we find that
 d 1 
ln  1 − 
dx  x + 1 x
lim ln f ( x ) = lim = − lim = −1
x →∞ x →∞ d x →∞ x + 1
dx
From continuity of logarithmic function we have

(
lim f ( x ) = −1 .
ln
x →∞ )
−1 1
Therefore, lim f ( x ) = e = . From this we can conclude that
x →∞ e
an +1 1
l = lim = <1
n →∞ an e
∞ n!
Hence, by Ratio test ∑ n
converges.
n =1 n
b) The sums of two series converge, if both of the sums converge. Hence the series
∞
(3n)!+ 4n +1 ∞
(3n)! ∞
4n +1
∑n =1 (3n + 1)!
will be convergent if ∑
n =1 (3n + 1)!
+∑
n =1 (3n + 1)!
convergent otherwise
∞
(3n)! 1 1 1 1
divergent. Since = and < , the series ∑ diverges by the p-
(3n + 1)! 3n + 1 6n 3n + 1 n =1 6 n
∞
(3n)! ∞
4n +1
series test. Thus we conclude that ∑
n =1 (3n + 1)!
diverges. On the other hand, ∑
n =1 (3n + 1)!
converges by ratio test, since

86
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

an +1 4n + 2 (3n + 1)! 4
lim = lim n +1
= lim = 0 <1.
n →∞ a n →∞ (3n + 2)! 4 n →∞ 3n + 2
n

∞
(3n)!+ 4n +1
This establishes that ∑
n =1 (3n + 1)!
diverges.

4n n ! n ! 4 n+1 (n + 1)!(n + 1)!

c) an = , an +1 = ,
(2n)! (2(n + 1))!
The ratio of the terms of the series is
an +1 4n +1 (n + 1)!(n + 1)! (2n)! 4(n + 1)(n + 1) 2(n + 1)
= ⋅ n = =
an (2n + 2)(2n + 1)(2n)! 4 n !n ! 2(n + 1)(2n + 1) 2n + 1
Computing limit we see that the ratio test fails to indicate convergence or divergence of the
given series. That means
a 2n + 2
lim n +1 = lim =1
n →∞ a n →∞ 2n + 1
n
a 2n + 2
However, n +1 = is always greater than 1and all the terms are greater than or equal
an 2n + 1
to a1 = 2 .
Thus
lim an ≠ 0
n →∞
∞
4n n ! n !
By contra positive of the n th term divergence test ∑
n =1 (2 n)!
is divergent.

1 ⋅ 3 ⋅ 5 ⋅⋯ (2n − 1) 1 ⋅ 3 ⋅ 5 ⋅⋯ (2n + 1)
d) Let an = , an +1 = n +1 n +1
n n
4 2 n! 4 2 (n + 1)!
Evaluating the limit yields,
a 1 ⋅ 3 ⋅ 5 ⋅⋯ (2n + 1) 4 n ⋅ 2n ⋅ n !
lim n +1 = lim n +1 n +1 ⋅
n →∞ a
n
n →∞ 4 ⋅ 2 ⋅ (n + 1)! 1 ⋅ 3 ⋅ 5 ⋅⋯ (2n − 1)
2n + 1
= lim
n →∞ 4 ⋅ 2( n + 1)

1
= <1
4
∞
1 ⋅ 3 ⋅ 5 ⋅⋯ (2n − 1)
By ratio test ∑n =1 4 n 2n n !
is convergent.

From the above examples we observe that the ratio test is especially useful when applied to a
∞
series ∑a
n =1
n in which the terms an involve powers or factorials. Also the Ratio works well

when an defined recursively. It will fail when an a rational function of n.

87
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

2.4.5 ROOT TEST

The Root test works well when an contains an expression to the nth power not factorials. This
is one reason why the Ratio test is more frequently used than the Root test.

Theorem5.14: Let ∑ an be a series of positive terms such that lim n an = ℓ

n →∞

(possibly l = ∞ )
i ) if 0 ≤ ℓ < 1, then ∑ an is convergent
ii ) if ℓ > 1, then ∑ an is divergent
iii ) if ℓ = 1 , the Root test fails to indicate convergence or divergence of a series.

Proof: Consult any reference book.

Example2.25 Test the convergence of the following series
2 −n
∞ nn  n + 1  n + 1 n 
∞
a) ∑ b) ∑    − 1
n2 
n =1 n  n 
n =1 ( n + 1) 
∞ n
 6n 
c) ∑  
n =1 2n − 1 
2
nn
Solution: a) an = . Taking nth root and computing its limit we get
n2
( n + 1)
1
 n2 n n
lim n an = lim  n  = lim  n  = 1
n →∞  2   
x →∞  n + 1 
n →∞  ( n + 1)n e
 
2
∞ nn
Therefore, by the Root test ∑ n2
is convergent.
n =1 ( n + 1)

88
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

−n
 n + 1  n + 1  n  
b) an =   − 1  and its nth root will be
 n  n   
 
n
−
 n + 1  n + 1   n n
n 1
n an =   − 1  .
 n   n 
=
 n + 1  n + 1 n
  
  −1
 n 
Now taking limit we obtain

lim
n →∞
( )
an = lim
n
n →∞ n + 1
.
 1
1
n
=
1
e −1
<1
lim 1 +  − 1
n →∞  n
−n
∞  n + 1  n + 1 n 
Since lim n an
n →∞
( ) =
1
e −1
< 1 , by root test ∑  
n =1  n  n 
 − 1 is convergent.

1
n n
 6n  n a =
6n  n 6n
c) Here an =   and n   =
 2 n − 1   2n − 1  2n − 1
6n
lim n an = lim = 3 >1
n →∞ n →∞ 2n − 1
∞ n
 6n 
Therefore, by Ratio Test ∑   is divergent.
n =1  2n − 1 
Remarks: The Ratio Test is frequently easier to apply than the Root Test, since it is usually
easier to compute ratio than nth roots. However, the Root Test has wider scope. Whenever the
Ratio Test shows convergence, the Root Test does too; whenever the Root Test is
inconclusive, the Ratio test is too.
Example2.26 Consider the series
1 1 1 1 1 1
+ + + + + +⋯
2 3 22 32 23 33
Does this series converge?
1 1 1
Solution: Here, a2 n −1 = n , a2 n = n and a2 n +1 = n +1
2 3 2
n
a 2n +1 2
lim 2 n = lim n = 2 lim   = 0
n →∞ a
2 n +1
n →∞ 3
 
n →∞ 3

and

89
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

n
a2 n−1 3n 3
lim = lim n = lim   = ∞
n →∞ a
2n
n →∞ 2
 
n →∞ 2

an
Thus, lim does not exist as we have different results for n even and odd.
n →∞
an +1
Therefore, the Ratio Test fails to indicate convergence or divergence of the given series.
On the other hand
1
1
 1  2n 1
lim(a2 n ) 2n
= lim  n  = <1
n →∞
 
n →∞ 3 3
and
1
 1  2 n +1
1
1
lim ( a2 n +1 ) = lim  n +1 
2 n +1 = < 1.
n →∞

n →∞ 2
 2
Thus, lim n an < 1 whether n is odd or even.
n →∞
∞
1 1 1 1 1 1
Therefore, by the Root Test, ∑a
n =1
n = + + + + + + ⋯ is convergent.
2 3 2 2 32 23 33
Exercise2.3
1. Test the convergence of the following series
∞ ∞ ∞ 2n
1 2k  2n 
a) ∑ b) ∑ 3 e) ∑  
n=2 ( ln n )n k =1 k n =1 n + 1 

3n − 2
∞ n ∞ ∞
n k
c) ∑   d)∑ k
f )∑ n
n =1  10  k =1 3 n =1 n
n
 2n , n odd ∞
2. Let an =  . Does ∑ an convergences?
 1 , n even n =1
 2 n

3. Prove that the series

2
∞
12 ⋅ 52 ⋅ 9 ⋅⋯ (4n − 3) 2
∑
n =1 3 ⋅ 7 ⋅11 ⋅⋯ (4n − 1)
2 2 2 2

is divergent.
4. Test whether the following series are convergent or not.
∞ ∞
(n !)n 1 ⋅ 3 ⋅ 5 ⋅⋯ (2n − 1)
a) ∑ n 2 b) ∑
n =1 [2 ⋅ 4 ⋅ 6 ⋅ ⋯ (2n )](4 + 3)
n
n =1 ( n )
∞
5an
c) ∑n =1
an , where a1 = 3 and an +1 =
2n + 1

90
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

5
 n 2 , if n is even a ∞
5) Let an =  . Verify that lim n +1 does not exist, but the series
→∞
∑a n
 2 , if n is odd an
n
n =1
 n 2

converges.

2.5 ALTERNATING SERIES

An arbitrary series can contain both positive and negative terms without any systematic
order. Such types of the series for which consecutive terms have opposite signs is said to be
alternating series. For more precisely we have the following definition.
Definition2.8 A series of the form
∞
n +1 n +1
∑ ( −1) an = a1 − a2 + a3 − a4 + ... + ( −1) an + ... or
n =1
∞
∑ ( −1) an = −a1 + a2 − a3 + a4 + ... + ( −1) an + ... Where
n n

n =1
an > 0 ∀n ∈ ℕ having an alternating negative and positive signs is called an alternating
series. Now we have the following result.

Theorem2.15 (Alternating series Test or Leibniz’s Theorem)

In the series
∞
n +1 n +1
∑ ( −1) an = a1 − a2 + a3 − a4 + ... + ( −1) an + ... or
n =1
∞
∑ ( −1) an = −a1 + a2 − a3 + a4 − ... − ( −1) an + ...
n n

n =1
if i) each an is positive ii) an ≥ an +1 for all n ≥ N , for some integer N and
∞ ∞
n +1
iii) lim an = 0 , then both the series ∑ ( −1) ∑ ( −1)
n
an and an converge.
x →∞ n =1 n =1

Example2.27: Determine whether the following series convergent or divergent

∞
a) ∑
( −1)n +1 b) ∑
∞ ( −1)n +1 ∞
c) ∑
( −1)n n d) ∑
(−1)n −1 ln n
∞

n =1 n n =1 n! n =1 n2 + 1 n=2 n
1 1
Solution: a) Here an = , and an +1 = . Then
n n +1
91
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

i) an is positive for each n

1 1
ii) > ∀n ∈ N implies an > an +1
n n +1
1
iii) lim an = lim =0.
n →∞ n →∞ n

∞ ( −1)n +1
Therefore, by alternate series test ∑ n
is convergent.
n =1
1 1
b) an = , an +1 = and
n! ( n + 1)!
i) an is positive for each n
1 1
ii) an ≥ an +1 ∀n ∈ N since >
n! ( n + 1)!
1
iii) lim an = lim =0
n →∞ n →∞ n!
∞ ( −1)n +1
Hence, ∑ n!
is convergent by alternating series test.
n =1

n x
c) an = 2 . Let f ( x ) = 2 for x ≥ 1 , then f (n) = an for all n ≥ 1 .
n +1 x +1
i) an is positive for each n
x2 + 1 − 2 x2 1 − x2
ii) f ' ( x ) = = ≤ 0∀x ≥ 1 which implies that f is decreasing on
( x + 1) ( x + 1)
2 2 2 2

[1, ∞ ) and consequently an is also decreasing on [1, ∞ ) .

n  1/ n 
iii) lim an = lim = lim  =0
n →∞ n →∞ n 2 + 1 n →∞  1 + 1/ n 2 
∞ n
n +1
Thus we can conclude that ∑ ( −1) is convergent by alternating series test.
n =1 n2 + 1
d) As in the above examples,
ln n
i) an = is positive for each n n ≥ 2 .
n

92
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

1
x( ) − ln x
ln x 1 − ln x
ii) Let f ( x) = , then f ′( x) = x 2 = . For x > 0, f ′( x) < 0 .
x x x2
Thus f ( x), and hence {an } is decreasing. Finally,
1
ln n
iii) lim an = lim = lim n = 0 by L’Hopital’s Rule.
n →∞ n →∞ n n →∞ 1
n −1
(−1) ln n
Therefore, ∑ converges.
n
Remark: All the tests studied previously can’t be applied to test the convergence of
alternating series except the nth term test.
∞
(−1) n n n
Example 2.28 The series ∑
n =1 n + 1
is an alternating series with lim an = lim
n →∞ n →∞ n + 1
= 1 ≠ 0 and

the alternating series tests fail to indicate convergence. But by the nth term test the given
series diverges.
∞
Closely examining the alternating behavior of the series ∑ (−1)
n =1
n +1
an we can determine a

bound for the difference between the n th partial sum and the sum of the series. The bound
for this difference is also a bound for the truncation error that occurs if Sn is used to
∞
approximate the actual sum S = ∑ (−1) n+1 an . Since the difference between the limit of the
n =1

series and the n th partial sum does not exceed an +1 , sn may be used to approximate s to

within an +1 . Consequently we have the following estimation theorem.

Theorem2.16: (The Alternating Series Estimation Theorem)

∞
n +1
If ∑ ( −1) an a convergent alternating series with 0 < an+1 < an for each integer n and
n =1

s is its sum, then

sn − s < an +1

for each natural number n.

93
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

∞ 1
n +1
Example 2.29 Find an approximation of ∑ ( −1) with an error less than 0.005.
n =1 n3
Solution: The partial sums sn will approximate the sum of the series with an error less than
0.005 that is,
1 1
sn − s < an+1 = < 0.005 =
(n + 1) 3
200

(n + 1)3 > 200

n + 1 > 3 200 ≃ 5.84

n > 4.84 .
Now we can choose n=5. If n = 5 , then the desired approximation is s ≃ s5 .
∞
(−1) n +1 5
(−1) n +1 1 1 1 1
∑
n =1 n 3
≃ s5 = ∑
n =1 n 3
= 1− 3 + 3 − 3 + 3
2 3 4 5
1 1 1 1
s5 = 1 − + − + ≃ 0.9044.
8 27 64 125
∞
(−1) n +1
Is it meant that ∑
n =1 n3
converges to the sum 0.9044 ?

π2 ∞
(−1) n +1
Example2.30 Approximate the sum of the series =∑
12 n =1 n2
with an error less than 0.01. Use the corresponding partial sum and error estimate to verify
that 3.113 < π < 3.15
Solution: The partial sums sn will approximate the sum of the series with an error less than

0.01 if an +1 < 0.01 that is

1
an +1 = < 0.01
(n + 1) 2

(n + 1) 2 > 100
n>9
If n = 10 , then the desired approximation is s ≈ s10 or
∞
(−1) n +1 1 1 1 1 1 1 1 1 1
∑
n =1 n 2
≈ s10 = 1 − + − + − + − + −
4 9 16 25 36 49 64 81 100

94
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

= 1.183865 − 0.3659027 = 0.8179622

∞
(−1) n +1
Therefore, ∑
n =1 n2
≈ 0.8179622 . Applying error estimation theorem for alternating series

implies
s − sn < 0.01
Solving the inequality we get
s10 − 0.01 < s < s10 + 0.01

Substituting for s10 and simplifying we obtain

0.8079622 < s < 0.8279622
π2
Again substituting for s = and simplifying we easily obtain the desired approximation.
12
3.1137672 < π < 3.1520702

95
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Exercise 2.4

1. Approximate ∑
∞ ( −1)n with an error less than 10-2
n =1 n!
∞ cos nπ
2. Approximate the sum ∑ n
with an error less than 10-3
n =1
(−1) n +1
∞
3. For what value of n is sn , the n th
partial sum, will approximate s = ∑ within
n =1 3n + 5
0.005?
4. a) The series
1 1 1 1 1 1 1 1
− + − + − +⋯ + n − n +⋯
3 2 9 4 27 8 3 2
fails to satisfy one of the conditions of the Alternating Series Test.
Which one?
b) Find the sum of the series in a).
5. Determine whether the following series converges or diverge.
∞
n3 ∞
sin(nπ / 2) ∞
n2
a) ∑ (−1)
n =1
n

n3 + en
b) ∑ (−1)
n =1
n

(n3 + n)1/ 2
c) ∑ (−1)
n =1
n

πn
∞
n+2 en ∞
1 nπ
d) ∑ (−1)
n =1
2n

n + 3n + 1
2
e) ∑ (−1) n −1

nπ n
f) ∑n
n =1
2
cos
2
∞
(3n)! ∞
1 nπ ∞
e− n
g) ∑ (−1)n
n =1 (n !)3
h) ∑
n =1 n
sin
2
∑ (−1)n
n =1 1 + e− n

96
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

2.6 ABSOLUTE AND CONDITIONAL CONVERGENCE

So far we have seen convergence tests for two types of series i.e. nonnegative term series and
alternating series. In this sub section we will see the convergence of general series and the
∞
relationship between a general series ∑a
n =1
n and the positive-term series of its absolute values

∑a
n =1
n = a1 + a2 + a3 + ⋯.

Definition 2.9 (Absolute convergence)

∞
A series ∑ an is said to be absolutely convergent if the series of the absolute values of
n =1
the terms
∞
∑ an = a1 + a2 + a3 + ⋯
n =1
converges.
Example2.31 Test absolute convergence for the following series
∞ ( −1)n +1 ∞ ( −1)n +1 ∞ sin n + cos n
a) ∑ 4
b) ∑ n −1
c) ∑
n =1 n n =1 e n =1 n5
∞ ( −1)n +1 ∞ 1 ∞ 1
Solution: a ) ∑ = ∑ 4 is a p – series with p = 4 >1 and hence ∑
n =1 n4 n =1 n n =1 n
4

∞ ( −1)n +1
converges. Consequently ∑ converges absolutely.
n =1 n4
n +1
∞ ( −1) ∞ 1 ∞ e
b) ∑ =∑ = ∑ is a convergent geometric series
n =1 en −1 n =1 e
n −1
n =1 e
n

∞ ( −1)n +1
Hence, ∑ converges absolutely
n =1 en −1
∞
sin n + cos n ∞
sin n + cos n
c) ∑ = ∑ . But cos n + sin n ≤ cos n + sin n ≤ 1 + 1 = 2 and
n =1 n5 n =1 n5
∞
sin n + cos n ∞
sin n + cos n ∞
sin n + cos n ∞
2
∑n =1 n 5
= ∑n =1 n 5
≤ ∑
n =1 n 5
≤ ∑
n =1 n
5

97
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

∞
2 ∞
sin n + cos n
Since ∑n
n =1
5
a convergent p-series with p = 5, ∑
n =1 n5
converges by the
∞
sin n + cos n
Comparison Test. Thus ∑ converges absolutely.
n =1 n5

Theorem 2.17: Every absolutely convergent series is convergent. That is

∞ ∞
If ∑ ak converges, then ∑ ak converges.
k =1 k =1

Proof: Exercise
∞ cos n
Example2.32 Show that ∑ 4
is convergent
k =1 n + 3n + 2

Solution: This is neither a positive term series nor an alternating series but does contain both
positive and negative terms. However,
cos n ∞ ∞
cos n 1 1 cos n 1
= ≤
n + 3n + 2 n + 3n + 2 n + 3n + 2 n
4 4 4
< 4
and ∑
n =1 n + 3n + 2
4
< ∑
n =1 n
4
.
∞
1
But ∑ 4 is a convergent p-series with p = 4>1. It follows from comparison test that
n =1 n
∞
cos n
∑
n =1 n + 3n + 2
4

converges.
∞ cos n
Therefore, ∑ 4
is absolutely convergent and hence convergent by the above
k =1 n + 3n + 2
theorem.
∞ ∞
Remark: The convergence of ∑ an does not necessarily imply that
n =1
∑a
n =1
n absolutely

convergent.
∞
(−1) n +1 ∞
(−1) n +1 ∞
1
Example2.33 ∑
n =1 n
is a convergent alternating series, but ∑
n =1 n
= ∑ is
n =1 n

divergent.
(−1) n +1
∞
Therefore, ∑ is not absolutely convergent
n =1 n
∞ ∞
Note: If ∑ an is divergent, then we cannot say any thing about ∑ an .
n =1 n =1

98
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

Definition2. 10 (Conditional Convergence)

∞ ∞ ∞
The series ∑ an is said to be conditionally convergent, if
n =1
∑ an is convergent and
n =1
∑a
n =1
n is

divergent.
Example 2.34 Determine whether the following series converges conditionally, converges
absolutely, or diverges.
∞ ∞
(−1) n (−1) n
a) ∑
n =1 ln( n + 1)
b) ∑
n =1 n
1 1
Solution: a) Let an = , then f ( x) = for x ≥ 1 and
ln(n + 1) ln( x + 1)
1
f ( n ) = an = n ≥ 1.
ln(n + 1)
i) an is positive for each n.
−1
ii) f ′( x) = < 0 for all x in [1, ∞) . Thus {an } is decreasing sequence for
( x + 1)(ln( x + 1))2
each n and
1
iii) lim an = lim =0
n →∞ n →∞ ln( n + 1)

∞
(−1) n
Hence, ∑
n =1 ln( n + 1)
converges by alternating series test.
∞
(−1) n ∞
1
On the other hand ∑
n =1
=∑
ln(n + 1) n=1 ln(n + 1)
and ln(n + 1) < n + 1 for each n ≥ 1 . As a result
∞ ∞
1 1
∑ n + 1 < ∑ ln(n + 1)
n =1 n =1
∞
1
But ∑ n + 1 divergent by p-series since p=1.
n =1
Implies
∞
1
∑ ln(n + 1)
n =1
∞
(−1) n
diverges by the comparison test . Thus ∑
n =1 ln( n + 1)
converges conditionally.

1
b) The series is alternating series with decreasing to 0 so it converges. But
n
∞
(−1) n ∞
1 1
∑
n =1 n
=∑
n =1 n
diverges (p-series with p= ).
2
∞
(−1) n
Hence, ∑
n =1 n
converges conditionally.

99
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

In this section we observed that the concept of convergence of an infinite series can be split
∞ ∞
Theorem2.18: If ∑ an
n =1
and ∑b
n =1
n converges absolutely and c is a constant, then
∞ ∞

∑
n =1
(an + bn ) and ∑
n =1
can are absolutely convergent.

into two separate sub concepts: absolute convergence and conditional convergence. Also
every series behave in exactly either, diverge, converge absolutely, or converge
conditionally.

100
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

2. 7 THE GENERALIZED CONVERGENCE TESTS

∞
Theorem2.19: Let ∑a
n =1
n be a series.
∞
a) Generalized Comparison Test: If an ≤ bn for n ≥ 1 , and if ∑b
n =1
n converges,
∞
then ∑a
n =1
n converges (absolutely)

∞
an
b) Generalized Limit Comparison Test. If lim
n →∞ bn
= l, l ∈ ℝ + and if ∑b
n =1
n

∞
converges, then ∑a
n =1
n converges (absolutely).
∞
c) Generalized Ratio Test: Suppose ∑a
n =1
n is a series of nonzero terms and

an +1
lim = l , where l represents either a nonnegative real number or infinity.
n →∞ an

i) If 0 ≤ l < 1, the series is absolutely convergent.

ii) If l > 1 , the series is divergent
iii) If l = 1 , no conclusion about convergence can be made. The series might
be absolutely convergent, conditionally convergent, or divergent.

d) Generalized Root Test: Suppose that lim n an = l , where l represents either a

n →∞

nonnegative real number or infinity.

i) If 0 ≤ l < 1, the series is absolutely convergent.
ii) If l > 1 , the series is divergent
iii) If l = 1 , then from this test alone we cannot draw any conclusion about
the convergence of the series.

101
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Example 2.35 Test the following for absolute convergence

∞ ∞
(−1) n (n + 3)! (−1) n n
a) ∑
n =1 3!n !3n
b) ∑
n =1 6n

(−1)n (n + 3)! (−1) n +1 (n + 4)! (−1)(−1) n (n + 4)(n + 3)!

Solution: a) Here an = , a n +1 = =
3!n !3n 3!(n + 1)!3n +1 3!(n + 1)n !3n (3)

an +1 (−1)(−1) n (n + 4)(n + 3)! 3!n !3n (−1)(n + 4) n+4

= ⋅ = =
an 3!(n + 1)n !3 (3)
n
(−1) (n + 3)!
n
3(n + 1) 3n + 3

Taking limit as n tends to infinity we get

an +1 n+4 1
lim = lim = < 1.
n →∞ an n →∞ 3n + 3 3
∞
(−1) n (n + 3)!
Therefore, by Generalized Ratio Test ∑ converges absolutely.
n =1 3!n !3n

(−1) n n (−1) n n n
n
b) an = and n an = n =
6n 6 n
6

n
n 1
lim n an = lim = <1
n →∞ n →∞ 6 6
∞
(−1) n n
Thus by Generalized Root Test ∑
n =1 6n
converges absolutely.

∞
n! xn
Example2.36 Show that the series ∑ n converges for x < e .
n =1 n

Solution: If x = 0 , trivially the given series converges. If x ≠ 0 , then

n
an +1 (n + 1)n ! x n ( x) n n  n 
= ⋅ =  x
an (n + 1) (n + 1) n ! x
n n
 n +1
And
n n
a  n   n  1
lim n +1 = lim   x = lim   x = x
n →∞ a
n
n →∞
 n +1  n →∞
 n +1 e

an +1 1
Thus the Generalized Ratio test implies that the given series converges if lim = x < 1.
n →∞ an e

102
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

∞
n! xn
Hence, the series ∑
n =1 n
n
converges for x < e . What happen if x = e or x = −e ?

Let us analyze the convergence if x = e or x = −e .

∞
n !e n
If x = e , then the series becomes ∑
n =1 n
n
. In this case

n
an +1 (n + 1)n !e n+1 n n  n 
= n +1
⋅ =  e
an (n + 1) n !e  n + 1 
n

n
an +1  n  1
lim = lim   e = lim e = 1.
 n +1
n →∞ →∞ →∞ n
an n n
 1
1 + 
 n

an +1
But the Generalized Ratio Test fails to indicate convergence or divergence since lim = 1.
n →∞ an

Let us use the Stirling’s formula and n th term divergence test to examine the convergence.
n
n !e n n e
n
an = n ≈ 2π n   ⋅ n = 2π n
n e n

lim an ≈ lim 2π n = ∞
n →∞ n →∞

∞
n! xn
Therefore, the series ∑
n =1 n
n
diverges when x = e . Similarly show that it is also divergent

for x = −e .
What can you say about the interval (-e, e) and the length e?
This question will be answered in the next chapter.

103
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

√ Check list

Mark a tick (√) against each of the following tasks that you can perform. I can

Give definition of an infinite series --------------------------------------

Give the difference between series and sequence------------------------

Define convergence of infinite series---------------------------------------

Check whether a given series is convergent or divergent by calculating limits

of partial sum----------------------------------------------------------

Find a formula for the sequence of partial sums and evaluate its limit-

State and proof theorem on geometric series ---------------------------

State and Proof the nth term divergence test-------------------------------

State and Prove integral test----------------------------------------------------

Estimate the sum of convergent infinite series that satisfy the hypothesis of

integral test by integral test remainder estimate theorem--------------

Test whether a given series converges or diverges by comparison test, limit

comparison test, root test, or ratio test -------------------------------

State and Proof alternating series theorem-------------------------------

Approximate the sum of convergent series by using alternating series estimate

theorem-----------------------------------------------------------------

Define absolute convergence, and conditional convergence---------------

Determine whether a series converges conditional, absolutely or diverges--

State Generalized test---------------------------------------------------------------

104
AU DEPARTMENT OF MATHEMATICS, May 2008
 Unit Two: Infinite Series

Summary
∞ Start
Given a series ∑a
n =1
n : Start
Use one of the following tests:
∞
1. Comparison or
∑a Diverges No Is lim an = 0? 2. Limit Comparison,
n n →∞
n =1 If the series with which
to compare is easily
Yes recognized;
3. Ratio;
∞ If an involves factorials
∑a
n =1
n Converges
Does
∞

∑a n and /or powers of n;

n =1 4. Root;
absolutely converge?
If an involves an nth
power;
Yes 5. Integral,
If f(n)= an , f decreasing
∞
and continuous, and f
∞ Is ∑a n a can be integrated.
∑a
n =1
n Diverges n =1
positive –term
If these tests are not conclusive,
the answer to this question
No
series? likely cannot be determined by
methods in this module.

Yes
Series to use for comparison:
∞
1
∞ ∞
1. ∑ n ( p − series )
p

∑a n Converges
Is ∑ an alternating with
n =1
n =1
a. converges if p>1
n =1
an +1 < an For all n? b. diverges if p ≤ 1 .
conditionally ∞
2. ∑ ar
n =1
n
( geometric series )

No a. converges if r <1
b. diverges if r ≥ 1.
No methods have been discussed that will
∞
enable us to determine whether ∑a
n =1
n

diverges or converges conditionally

105
AU DEPARTMENT OF MATHEMATICS, May 2008
 UNIT TWO INFINITE SERIES

Exercise 2.5
1. Determine whether the following series absolutely convergent, conditionally
convergent, or divergent.
∞
(−1) n +1 ∞
1 ∞
1
a) ∑
n =1 (ln n)
n
b) ∑ (−1)n n sin
n =1 n
c) ∑ (−1)
n=2
n

n ln n
∞  1  n +1  1  n  ∞
(−1)n n 2 + 3
d) ∑ (−1) 1 +  − 1 +  
n
d) ∑ 3n 2
+ 2n − 1
n =1  n   n   n =1

∞
∞
1 ⋅ 3 ⋅ 5 ⋅⋯ (2n + 1) n 2 − 4n + 4 ∞
 3 
e) ∑ (−1)n
n =1 2 ⋅ 5 ⋅ 8 ⋅⋯ (3n + 2)
f) ∑
n =1 9n + 3n − 2
2
g) ∑ (−1)
n =1
n
 
 4n 
∞ ∞
∞
arctan (−1) n (1 + 4n) (−1) n
h) ∑
n =1 n + 1
2
i) ∑
n =1 7n 2 − 1
j) ∑
n = 2 ln(ln n )

∞
sin(nπ / 2) ∞
cos(nπ / 4) ∞
1.3.5⋯ (2n − 1)
k) ∑ (−1)n
n =1 (n3 + n)1/ 2
l) ∑
n =1 n2
o) ∑ (−1)
n =1
n +1

2 ⋅ 4 ⋅ 6⋯ (2n)
2. For the following find an upper bound for the error of the sum of the first five terms of the
series as an approximation of the sum of the series.
∞ ∞
n 1
a) ∑
n =1
(−1)
n!
n
b) ∑ (−1)
n=2
n +1

1 + n + 6n 2
3. Find the sum of the following telescoping series
∞
2n + 1 ∞
 1 1 
a) ∑
n =1 n (2 n + 1)
2 2
b) ∑  ln(n + 2) − ln(n + 1) 
n =1  
∞
c) ∑ (tan
n =1
−1
(n) − tan −1 (n + 1))

n −1
4. Suppose {an } is defined by a1 = 1 and an = an −1 if n ≥ 2 .
n
a) use the completeness property to show that {an } is convergent.

b) find an explicit formula for an

c) determine lim an
n →∞

5. For the following series find an approximation to the sum that is accurate to within 10−5 .
∞
(−1)n −1 ∞
(−1)n +1
a) ∑
n = 2 4 ( n − 1)
n
b) ∑
n =1 (2 n − 1)
3

106
AU DEPARTMENT OF MATHEMATICS, May 2008