MA2311

Chapter 2
Infinite Series

TAN BAN PIN

National University of Singapore

TAN BAN PIN MA2311 1 / 86

Overview

Infinite Series
Partial Sum
Convergence of Geometric Series
Telescoping Series
nth Term Test
p-series Test
Comparison Test
Limit Comparison Test
Ratio Test
Root Test
Alternating Series Test
Absolute Convergent/Conditionally Convergent/Divergent

TAN BAN PIN MA2311 2 / 86

Geometric Sequences (Progressions)

Examples

Consider a segment of length 1.

Cut half on the first day.
Cut half of the remaining in the second day.
In general, cut half of the remaining everyday.
How much have we cut by the 100th day?
100
1 1 1 1 1
X 1
2 + 2 2 + 2 3 + 2 4 + · · · + 2 100 = i
i=1
2

TAN BAN PIN MA2311 3 / 86

Geometric Sequences (Progressions)
Examples

Consider an 8 × 8 chessboard.
Put 1 grain of rice in the first square of the chessboard.
Doubling the number in the next square.
How much rice do we need to fill the chessboard?
X64
1 + 2 + 22 + 23 + 24 + · · · + 263 = 2i−1
i=1

TAN BAN PIN MA2311 4 / 86

Infinite Series

When we look at an infinite sequence {an }, we are looking at an infinite

list of numbers:
a1 , a2 , · · · , an , · · ·
If we consider the sum

a1 + a2 + a3 + · · · + an + · · · ,

we get an infinite series.

In notation, we write
∞
X
a1 + a2 + a3 + · · · + an + · · · = an .
n=1

The term an is called the n-th term of the series.

TAN BAN PIN MA2311 5 / 86

Infinite Series

From the sequence

1 1 1 1 1 1
, , , , , ,···
2 4 8 16 32 64
We obtain the infinite series,
1 1 1 1 1 1
+ + + + + + ···
2 4 8 16 32 64
1
The n-th term is an = 2n .

1 1 1
, , ,···
2 22 23
a1 , a2 , a3 , · · · , an , · · ·

TAN BAN PIN MA2311 6 / 86

Note

Infinite Sequence: a1 , a2 , a3 , · · · , an , · · ·
∞
X
Infinite Series: an = a1 + a2 + a3 + · · · + an + · · ·
n=1

TAN BAN PIN MA2311 7 / 86

Infinite Series

Pause and Think !!!

A bit of Advance Mathematics:

1 − 1 + 1 − 1 + 1 − 1 + 1 − 1 + · · · =?

Which of the following is true?

(a) (1 − 1) + (1 − 1) + (1 − 1) + (1 − 1) + · · ·
= 0 + 0 + 0 + 0 + · · · = 0.
(b) 1 + (−1 + 1) + (−1 + 1) + (−1 + 1) + · · ·
= 1 + 0 + 0 + 0 + · · · = 1.

TAN BAN PIN MA2311 8 / 86

Infinite Series

∞
X
Infinite Series: an = a1 + a2 + a3 + · · · + an + · · ·
n=1

What is the value of

a1 + a2 + a3 + · · · + an + · · · =?

TAN BAN PIN MA2311 9 / 86

Infinite Series

∞
X
Infinite Series: an = a1 + a2 + a3 + · · · + an + · · ·
n=1

What is the value of

a1 + a2 + a3 + · · · + an + · · · =?

We are adding infinite number of terms, so the sum might not be finite.
If the sum is finite, we may or may not be able to find the finite sum.
We want to know when an infinite series gives a finite sum.
If we get a finite sum, we say the infinite series is convergent, otherwise,
we say the infinite series is divergent.

TAN BAN PIN MA2311 10 / 86

Partial Sums
∞
X
Infinite Series: an = a1 + a2 + a3 + · · · + an + · · ·
n=1
Consider:

s1 = a1
s2 = a1 + a2
s3 = a1 + a2 + a3
..
.
sn = a1 + a2 + a3 + · · · + an
..
.

s1 , s2 , s3 , · · · , sn , · · · is called the sequence of partial sums.

The term sn is called the n-th partial sum.

TAN BAN PIN MA2311 11 / 86

Partial Sum

The nth partial sum of {an } is

n
X
sn = a1 + a2 + · · · + an = ak
k=1

the sum of its first n terms

TAN BAN PIN MA2311 12 / 86

Given an infinite series
∞
X
an = a1 + a2 + a3 + · · · + an + · · ·
n=1

we add ’all’ the terms of the sequence a1 , a2 , a3 , · · · , an , · · ·

So, we are adding infinite number of terms.

Consider the partial sum sn = a1 + a2 + a3 + · · · + an .

If we let n → ∞, we have

lim sn = a + a2 + · · ·
n→∞ |1 {z }
add infinite number
of terms since n → ∞

∞
X
Thus, we can think that an is the same as lim sn .
n→∞
n=1

TAN BAN PIN MA2311 13 / 86

Consider the partial sum sn = a1 + a2 + a3 + · · · + an .
If we let n → ∞, we have

lim sn = a + a2 + · · ·
n→∞ |1 {z }
add infinite number
of terms since n → ∞

∞
X
Thus, we can think that an is the same as lim sn .
n→∞
n=1

So if lim sn = L, then we can say that

n→∞

a1 + a2 + · · · + an + · · · = L
∞
X
an = L
n=1

TAN BAN PIN MA2311 14 / 86

If the sequence of partial sums {sn } is such that sn → L, then we say that
∞
X ∞
X
the series an is convergent and an = L (we say that the sum is L).
n=1 n=1
If the sequence of partial sums {sn } does not converge, then we say that
∞
X
the series an is divergent.
n=1

TAN BAN PIN MA2311 15 / 86

If the sequence of partial sums {sn } does not converge, then we say that
∞
X
the series an is divergent.
n=1

If the sequence of partial sums {sn } does not converge, it could be

because:
(i) lim sn does not exist, or
n→∞
(ii) lim sn = ∞ or −∞.
n→∞

TAN BAN PIN MA2311 16 / 86

If the sequence of partial sums {sn } does not converge, then we say that
∞
X
the series an is divergent.
n=1

If the sequence of partial sums {sn } does not converge, it could be

because:
∞ X
(i) an does not exist, or
n=1
∞
X
(ii) an = ∞ or −∞.
n=1

TAN BAN PIN MA2311 17 / 86

Pause and Think !!!
A bit of Advance Mathematics:

1 − 1 + 1 − 1 + 1 − 1 + 1 − 1 + · · · =?

s1 = 1
s2 = 1 − 1 = 0
s3 = 1 − 1 + 1 = 1
s4 = 1 − 1 + 1 − 1 = 0

∴ the sequence of partial sums alternates between 0 and 1.

So lim sn does not exist.
n→∞
Thus the infinite sum

1 − 1 + 1 − 1 + 1 − 1 + 1 − 1 + ···

is divergent.
TAN BAN PIN MA2311 18 / 86
Infinite Series

n
X
sn = a1 + a2 + · · · + an = ak
k=1

The set of partial sums, sn is itself a sequence.

Q. Does sn converge to a finite real number as n → ∞? That is, we
investigate whether the following exists:
n
X ∞
X
lim ak = ak
n→∞
k=1 k=1

TAN BAN PIN MA2311 19 / 86

Infinite Series

n
X ∞
X
We investigate if lim ak = ak
n→∞
k=1 k=1

∞
X
ak is the infinite series corresponding to the sequence {an }
k=1
We say that this infinite series is convergent if lim sn exists. Otherwise,
n→∞
it is divergent.
Thus, to check whether a given infinite series is convergent, we consider
the partial sum sn and determine whether lim sn exists or not.
n→∞

TAN BAN PIN MA2311 20 / 86

Geometric Series

For a 6= 0, the series

∞
X
a + ar + ar2 + · · · + arn−1 + · · · = arn−1
n=1

is called a geometric series, where a and r are fixed numbers,

a is called the first term and r is the (common) ratio .

TAN BAN PIN MA2311 21 / 86

Geometric Series

For this series, the n-th partial sum sn is given by

sn − rsn = a − arn
1 − rn
sn = a r 6= 1
1−r

TAN BAN PIN MA2311 22 / 86

Partial Sum

Example
Geometric Sequence with first term = a and common ratio = r

a(1 − rn )
sn =
1−r

TAN BAN PIN MA2311 23 / 86

 a + ar + ar2 + · · · + arn−1 + · · ·

(i) r = 1 a + a + a + a + ···
Then sn = na → ∞ if a > 0 (or −∞ if a < 0)
Thus, the series is divergent.

(ii) r = −1 a − a + a − a + ···
Then {sn } is a, 0, a, 0, · · ·
Thus, the series is divergent.

TAN BAN PIN MA2311 24 / 86

 1 − rn
sn = a
1−r

(iii) If |r| < 1, then rn → 0.

a
Thus, sn → 1−r .
a
Hence, the sum of the series is 1−r .

(iv) If |r| > 1, then rn → ∞ (or −∞), and the series diverges.

TAN BAN PIN MA2311 25 / 86

Convergence of Geometric Series

The geometric series

a + ar + ar2 + · · · + arn−1 + · · ·

with a 6= 0 converges to the sum

a
if |r| < 1
1−r

and
it diverges if |r| > 1.

TAN BAN PIN MA2311 26 / 86

Geometric Sequences (Progressions)
Examples

Consider an 8 × 8 chessboard.
Put 1 grain of rice in the first square of the chessboard.
Doubling the number in the next square.
How much rice do we need to fill the chessboard?
X64
1 + 2 + 22 + 23 + 24 + · · · + 263 = 2i−1
i=1

TAN BAN PIN MA2311 27 / 86

Geometric Sequences (Progressions)

Examples

Consider a segment of length 1.

Cut half on the first day.
Cut half of the remaining in the second day.
In general, cut half of the remaining everyday.
How much have we cut by the 100th day?
100
1 1 1 1 1
X 1
2 + 2 2 + 2 3 + 2 4 + · · · + 2 100 = i
i=1
2

TAN BAN PIN MA2311 28 / 86

Sum of Geometric Sequences

Examples
Consider a segment of length 1.
Cut half on the first day.
Cut half of the the remaining everyday.
How much have we cut by the 100th day?
Recall that first term a = 21 , common ratio r = 12 .
a
sn = −
1−r (1 h rn ).
1
i
1 100 1


s100 = 2
1− 12
1− 2 =1− 2100
1
In general, sn = 1 − 2n
As n gets larger, sn gets close to 1.

TAN BAN PIN MA2311 29 / 86

Example

1
(i) 19 + 27 1
+ 81 + · · · is a geometric series
first term a = 19 and common ratio r = 13 .
1
a
It converges to = 9 1
1−r 1− 3
1
= .
6

TAN BAN PIN MA2311 30 / 86

Example

(ii) 4 − 2 + 1 − 21 + 14 − · · ·
first term a = 4 and common ratio r = − 12 .
2
1 1 a
  
4+4 − +4 − + ··· =
2 2 1−r
4
=  
1 − − 12
8
=
3

TAN BAN PIN MA2311 31 / 86

Example (A convergent series)

∞  k
X 3
k=1
4
This is a geometric series, first term 34 , common ratio = 3
4

a(1−rn )
Geometric series first term a, common ratio r, sn = 1−r

3
(1−( 34 )n ) 3
(1−0)
Partial sum sn = 4
1− 34
→ 4
1− 43
3
sn → 4
1− 43
= 3 when n → ∞
Conclusion: The given infinite series is convergent

TAN BAN PIN MA2311 32 / 86

Infinite Series

∞
X
The geometric series ark−1 converges iff −1 < r < 1.
k=1
Furthermore, for −1 < r < 1,
∞
X a
ark−1 =
k=1
1−r

TAN BAN PIN MA2311 33 / 86

An Example on Geometric Series

Determine the range of values of x for which the series

∞
X 100xk
k=1
52k−1

is convergent and find its value in terms of x when it is convergent.

TAN BAN PIN MA2311 34 / 86

An Example on Geometric Series

Determine the range of values of x for which the series

∞
X 100xk
k=1
52k−1

is convergent and find its value in terms of x when it is convergent.

∞
X
Any geometric series takes the form Ark for some constants
k=1
A and r, where r is the common ratio.

100xk k x k
= 100 25kx·5−1 = 500


Observe that 52k−1 25 .

TAN BAN PIN MA2311 35 / 86

An Example on Geometric Series
Determine the range of values of x for which the series
∞
X 100xk
k=1
52k−1

is convergent and find its value in terms of x when it is convergent.

∞ ∞ k
100xk x
X X 
= 500
k=1
52k−1 k=1
25

x


We have a geometric series with first term = 500 25 = 20x and common
x
ratio 25 . Hence, this series converges if and only if
x
−1 < 25 < 1 ⇒ −25 < x < 25 . Furthermore, it converges to
20x 500
x =
1 − 25 25 − x

TAN BAN PIN MA2311 36 / 86

A Telescopic Series

∞
X 1 1 1 1
= + + + · · · =?
k=1
k(k + 1) 1·2 2·3 3·4
1 1 1
Solution: Use partial fraction” techniques to write k(k+1) = k − k+1 .
Then
1 1 1 1
sn = + + + ··· +
1·2 2·3 3·4 n · (n + 1)
1 1 1 1 1 1 1 1
= ( − ) + ( − ) + ( − ) + ··· + ( − )
1 2 2 3 3 4 n n+1
1 1 1 1 1 1 1
= 1 + (− + ) + (− + ) + · · · + (− + ) −
2 2 3 3 n n n+1
1
=1− , and
n+1
∞
X 1 1
= lim (1 − ) = 1.
k=1
k(k + 1) n→∞ n+1
TAN BAN PIN MA2311 37 / 86
Theorem

∞
X ∞
X
1. If the series an and bn are convergent, then the series
n=1 n=1
∞
X
(an + bn ) is also convergent and
n=1

∞
X ∞
X ∞
X
(an + bn ) = an + bn .
n=1 n=1 n=1

∞
X ∞
X
2. If the series an is convergent and c ∈ R, then the series can is
n=1 n=1
also convergent and
∞
X ∞
X
can = c an .
n=1 n=1

TAN BAN PIN MA2311 38 / 86

Infinite Series

n
X
sn = a1 + a2 + · · · + an = ak
k=1

The set of partial sums, sn is itself a sequence.

Q. Does sn converge to a finite real number as n → ∞? That is, we
investigate whether the following exists:
n
X ∞
X
lim ak = ak
n→∞
k=1 k=1

TAN BAN PIN MA2311 39 / 86

Infinite Series

n
X ∞
X
lim ak = ak
n→∞
k=1 k=1

∞
X
ak is convergent iff lim sn exists
n→∞
k=1
However, partial sums are often difficult to find.

TAN BAN PIN MA2311 40 / 86

Infinite Series

For geometric series and telescoping series, we are able to find sn in terms
of n, and hence lim sn . Thus, we are able to decide whether the given
n→∞
geometric/telescopic series is convergent or not.

For most series, we might not be able to find sn in terms of n.

There are methods for testing convergence of infinite series that do not
require the partial sums.

These methods are known as convergence tests for infinite series.

TAN BAN PIN MA2311 41 / 86

Infinite Series

A list of convergence tests:

nth term test (Test for divergence only)
p--series test (Applicable to p-series only)
Ratio Test
Root Test
Integral Test
Alternating Series Test
Comparison Test
Limit Comparison Test
· · · and many more · · ·

TAN BAN PIN MA2311 42 / 86

Infinite Series

nth term test (Test for divergence only)

If {an } does not converge to 0, then the infinite series
∞
X
ak
k=1

is divergent.

Intuitively, for an infinite series to be convergent, the terms begin summed

should become very small eventually, in the sense an should → 0.

TAN BAN PIN MA2311 43 / 86

Infinite Series

nth term test DOES NOT test for convergence

{an } converges to 0 does NOT necessarily imply that the series
∞
X
ak converges.
k=1

Intuitive Explanation
an → 0 alone is not enough to guarantee convergence of the series
because we need an to be ”small enough”. For example, both n1 and
1
n2
→ 0 but n12 is much smaller than n1 when n is large.

TAN BAN PIN MA2311 44 / 86

Proof of nth term test (Divergence Test)

∞
X
We will show ak converges ⇒ lim an = 0
n→∞
k=1
∞
X
Proof: Let ak = S.
k=1
The idea behind the proof is to use the facts that
n
X
1. S = lim sn where sn = ak .
n→∞
k=1
2. S = lim sn−1 as well, which is pretty obvious.
n→∞
3. an = sn − sn−1 .
All we have to do is to take limit on both sides of (3) to get

lim an = S − S = 0. (QED)
n→∞

TAN BAN PIN MA2311 45 / 86

Infinite Series

p-series test (Applicable to p-series only)

The series ∞
X 1
k=1
kp
(known as p-series) is convergent of p > 1 and is divergent of p 6 1

Intuitive Explanation To ensure convergence of the series, we need an to

be ”small enough”. For example, both n1 and n12 → 0 but n12 is much
∞
1
X 1
smaller than n when n is large. According to the above result, 2
k=1
k
∞
X 1
(p = 2) is convergent while (p = 1) is divergent.
k=1
k

TAN BAN PIN MA2311 46 / 86

Proof of p-series Test

Convergence of the p-series for p > 1,

∞
X 1
n=1
np

Sum of areas of rectangles over x6n 

 1 6<0
N Z N z }| {
X 1 1 1 N (1 − p) − 1

< dx =
n=2
np 1 x p 1−p  

Letting N → ∞, we conclude
∞ ∞
X 1 1 X 1
p
< ⇒ p
(converges)
n=2
n p − 1 n=1
n
(since it is smaller than a finite value)

TAN BAN PIN MA2311 47 / 86

Proof of p-series Test

Divergence of the p-series for p < 1,

∞
X 1
n=1
np

Sum of areas of rectangles over x6n 

 1 6>0
N Z N z }| {
X 1 1 1 N (1 − p) − 1

> dx =
n=1
np 1 x p 1−p  

Letting N → ∞, we conclude
∞
X 1
p
= ∞ (diverges)
n=1
n
(since RHS → ∞)

TAN BAN PIN MA2311 48 / 86

∞
X 1
(The Harmonic Series) is said to be divergent. Why?
n=1
n
A quick proof
2 terms 4 terms 8 terms
∞
z }| { z }| { z }| {
X 1 1 1 1 1 1 1 1 1 1 1
= 1 + + ( + )+( + + + )+( + + ··· + )+···
n=1
n 2 3 4 5 6 7 8 9 10 16
2 terms 4 terms 8 terms
z }| { z }| { z }| {
1 1 1 1 1 1 1 1 1 1
> 1 + + ( + )+( + + + )+( + + ··· + )+···
2 4 4 8 8 8 8 16 16 16
1 1 1
= 1 + + + + ···
2 2 2
1 1 1
RHS = 1 + + + + · · · diverges to infinity
2 2 2
Hence, so does LHS.

TAN BAN PIN MA2311 49 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.

Recall nth term test: If an does not converge to 0, then the infinite
∞
X
series ak is divergent.
k=1

TAN BAN PIN MA2311 50 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.
Check that
2n + 1 2n + 1 1 2 1 2
     
lim 2 = lim lim 1 lim 1
n→∞ n→∞ 3n + 2 n→∞ n→∞
(3n + 2)(5n) n 5n nn

Applying limit laws: lim an = A, lim bn = B, lim (an bn ) = AB

n→∞ n→∞ n→∞

TAN BAN PIN MA2311 51 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.
Check that
2n + 1 (2n + 1)/n 1 2 1 2
     
lim 2 = lim lim 1 lim 1
n→∞ n→∞ (3n + 2)/n n→∞ n→∞
(3n + 2)(5n) n 5n nn

Divide by highest power, n

TAN BAN PIN MA2311 52 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.
Check that
1
! 2 2
2n + 1 2+ 1 1
 
n
lim 2 = lim 2 lim 1 lim 1
n→∞ n→∞ 3+ n→∞ n→∞
(3n + 2)(5n) n n 5n nn

Apply limit law: lim f (an ) = f ( lim an ), where f (x) = x2

n→∞ n→∞

TAN BAN PIN MA2311 53 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.
Check that
2 + n1
!
2n + 1 1 2 1 2
  
lim 2 = lim lim 1 lim 1
n→∞ n→∞ 3 + 2 n→∞ n n→∞ n
(3n + 2)(5n) n   n 5 n
= 32 (1)2 (1)2 = 32

TAN BAN PIN MA2311 54 / 86

Example

Use the nth term test to show that the series

∞
X 2k + 1
2
k=1 (3k + 2)(5k) k

is divergent.
Now,
2n + 1 2
lim 2 = 6= 0
n→∞
(3n + 2)(5n) n 3

Recall nth term test: If an does not converge to 0, then the infinite
∞
X
series ak is divergent.
k=1

Hence, by the nth term test, the series is divergent.

TAN BAN PIN MA2311 55 / 86

Definition

∞
X
A series ak is called an (eventually) non-negative series if each
k=1
ak > 0 (eventually).
Here ak > 0 eventually means there exists K ∈ N such that ak > 0 for all
k > K (for example, ak > 0 for all k > 15).
We will first develop tests for (eventually) non-negative series.

TAN BAN PIN MA2311 56 / 86

Theorem (Comparison Test)
∞
X ∞
X
Consider two (eventually) non-negative series ak and bk . Suppose
k=1 k=1
that there exists K ∈ N such that

0 6 ak 6 bk for all k > K.

∞
X ∞
X
1. If bk converges, then ak converges.
k=1 k=1
X∞ ∞
X
2. If ak diverges, then bk diverges.
k=1 k=1
[Roughly speaking, the comparison Test should be understood as follows:
Suppose we have two (eventually) non-negative series such that one series
is (eventually) termwise bigger than the other series.
1. If the bigger series has a finite sum, then so does the smaller series.
2. If the smaller series has an infinite sum, then so does the bigger series.]

TAN BAN PIN MA2311 57 / 86

Examples on Comparison Test

∞
X 1
1. Is the series convergent?
n=1
3n +2
Solution: Note that
1 1
06 6 n for all n ∈ N,
3n +2 3
∞
X 1 1
and converges because it is geometric series with |r| = 3 < 1. So
n=1
3n
∞
X 1
by Comparison Test, the series converges.
n=1
3n +2

TAN BAN PIN MA2311 58 / 86

Examples on Comparison Test

∞
X 4 + cos n
2. Is the series convergent or divergent?
n2
n=1
Solution: For all n > 1, we have

−1 6 cos n 6 1

⇒ 3 6 4 + cos n 6 5
3 4 + cos n 5
⇒06 2 6 2
6 2.
n n n
∞ ∞
X 5 X 1
Now, =5 is convergent, since it is a p-series with p = 2 > 1.
n=1
n2 n=1
n2
∞
X 4 + cos n
Thus, by the Comparison Test, also converges.
n=1
n2

TAN BAN PIN MA2311 59 / 86

Theorem (Limit Comparison Test)
∞
X ∞
X
Let an and bn be two eventually positive series, and suppose that
n=1 n=1
the limit
an
ρ = lim exists.
n→∞ bn
1. If ρ > 0, then either the two series both converge or both diverge.
∞
X ∞
X
2. If ρ = 0 and bn converges, then an converges.
n=1 n=1
[Roughly speaking, in Case 1, the two series resemble finite non-zero
multiple of each other. Therefore, either both series have finite sum or
both series have infinite sum. ∞ X
In Case 2, the condition ρ = 0 roughly means that an is smaller than
n=1
∞
X
bn , so the conclusion is similar to that of the Comparison Test.]
n=1

TAN BAN PIN MA2311 60 / 86

Examples on Limit Comparison Test

∞
X1
1. Is the series 2−n+1
convergent?
n=1
n
∞
X 1
Solution: We know that 2
is a p-series with p = 2 > 1, so it
n=1
n
∞
X 1
converges. It seems reasonable to compare the given series with ,
n=1
n2
1
but unfortunately n2 −n+1 > n12 .
So comparison test fails. We use the limit comparison test instead.

TAN BAN PIN MA2311 61 / 86

Examples on Limit Comparison Test
∞
1
X
1. Is the series 2−n+1
convergent?
n=1
n
Solution (Cont’d): We use the limit comparison test instead:
1
n2 − n + 1 1 1
 
n2
ρ= lim = lim = lim 1 − + 2 = 1 > 0.
n→∞ 2 1 n→∞ n 2 n→∞ n n
n −n+1

∞
X 1
So by the Limit Comparison Test, either the two series and
n=1
n2
∞
X 1
both converge or both diverge.
n=1
n2 −n+1
∞
X 1
Since converges (being a p-series with p = 2 > 1), so
n=1
n2
∞
X 1
2
also converges.
n=1
n −n+1
TAN BAN PIN MA2311 62 / 86
Examples on Limit Comparison Test

∞
1
X
2. Is the series convergent?
n=1
n+2
1
Solution: Note that n+2 > 0 for each n ∈ N.
1
n+2 2
 
ρ= lim n1 = lim = lim 1 + = 1 > 0.
n→∞
n+2
n→∞ n n→∞ n
∞
X 1
So by the Limit Comparison Test, either the two series and
n=1
n
∞ ∞
X 1 X 1
both converge or both diverge. Since diverges (being a
n=1
n+2 n=1
n
∞
X 1
p-series with p = 1), so also diverges.
n=1
n + 2

TAN BAN PIN MA2311 63 / 86

Theorem (Ratio Test)

∞
X
Let an be an eventually positive series, and suppose that the limit
n=1

an+1
ρ = lim exists.
n→∞ an
∞
X
1. If ρ < 1, then the series an converges.
n=1
X∞
2. If ρ > 1, then the series an diverges.
n=1
3. No conclusion if ρ = 1. (Why?)

TAN BAN PIN MA2311 64 / 86

Theorem (Ratio Test)
∞
X
Let an be an eventually positive series, and suppose that the limit
n=1
an+1
ρ = lim exists.
n→∞ an
∞
X 1
3. The case ρ = 1. The series diverges, and
n=1
n
1
n+1 n
ρ= lim = lim = 1.
n→∞ 1 n→∞ n +1
n
∞
X 1
On the other hand, the series converges, and
n=1
n2
1 2
n

(n+1)2
ρ = lim 1 = lim = 1.
n→∞
n2
n→∞ n+1
∞
X
Hence when ρ = 1, an may converge or diverge, and no conclusion can
n=1
be drawn.
TAN BAN PIN MA2311 65 / 86
Examples on Ratio Test

∞ n
X 2
1. Is the series convergent or divergent?
n!n=1
n
Solution: For each n ∈ N, an = 2n! > 0. And

2n+1
an+1 (n+1)! 2
ρ = lim = lim 2n = lim = 0 < 1.
n→∞ an n→∞ n→∞ n+1
n!

∞ n
X 2
Thus by the Ratio Test, the series is convergent.
n=1
n!

TAN BAN PIN MA2311 66 / 86

Examples on Ratio Test
∞
X (2n)!
2. Is the series convergent or divergent?
n=1
(n!)2
(2n)!
Solution: For each n ∈ N, an = (n!)2
> 0. And

(2(n+1))!
an+1 [(n+1)!]2
ρ = lim = lim (2n)!
n→∞ an n→∞
(n!)2
(2n + 2)!(n!)2
= lim
n→∞ (2n)![(n + 1)!]2

(2n + 2)(2n + 1)
= lim = 4 > 1.
n→∞ (n + 1)2
∞
X (2n)!
Thus by the Ratio Test, diverges.
n=1
(n!)2

TAN BAN PIN MA2311 67 / 86

Examples on Ratio Test
∞
X
n
3. Is the series an , where a1 = 1 and an+1 = 2n+1 an , convergent or
n=1
divergent?
n an+1 n
Solution: From an+1 = 2n+1 an , we have an = 2n+1 .

an+1 n
ρ = lim = lim
n→∞ an n→∞ 2n + 1
n
n
= lim
n→∞ 2n
+ n1
n
1
= lim
n→∞ 2 + 1
n
1
= < 1.
2
∞
X
By the Ratio Test, an converges.
n=1

TAN BAN PIN MA2311 68 / 86

Examples on Ratio Test
∞
X 3n
4. Determine if is convergent.
n=1
2n + 5
Solution:
an+1 3n+1 2n + 5
ρ = lim = lim n+1 ·
n→∞ an n→∞ 2 +5 3n
n
2 +5
= 3 lim n+1
n→∞ 2 +5
5
!
1 + 2n Divide both numerator
= 3 lim
n→∞ 2 + 5n
2
and denominator by 2n
1 5
   
=3 → 0 as n → ∞
2 2n
3
= >1
2
∞
X 3n
By the Ratio Test, is divergent.
n=1
2n + 5
TAN BAN PIN MA2311 69 / 86
Theorem (Root Test)

∞
X
Let an be an eventually non-negative series, and suppose that the limit
n=1

1
ρ = lim ann exists.
n→∞

∞
X
1. If ρ < 1, then the series an converges.
n=1
X∞
2. If ρ > 1, then the series an diverges.
n=1
3. No conclusion if ρ = 1.

TAN BAN PIN MA2311 70 / 86

Example on Root Test

∞
X [2(n + 1)]n
Is the series convergent or divergent?
n=1
nn+1
[2(n+1)]n
Solution: Note that each n ∈ N, an = nn+1
> 0. Also,
1
[2(n + 1)]n n
1

n
ρ = lim an = lim
n→∞ n→∞ nn+1
2 · (n + 1)
= lim 1
n→∞
n1+ n
2 n+1
= lim 1 · = 2 > 1.
n→∞ n
n n

Thus by the Root Test, the series diverges.

TAN BAN PIN MA2311 71 / 86

Definition

An alternating series is a series of the form

∞
X
(−1)k+1 ak = a1 − a2 + a3 − a4 + · · · , or
k=1

∞
X
(−1)k ak = −a1 + a2 − a3 + a4 − · · ·
k=1

with each ak > 0.

Examples
∞
X (−1)n+1 1 1 1
1. = 1 − + − + ···
n=1
n 2 3 4
∞ n
X (−1) 1 1 1
2. 2
= −1 + 2 − 2 + 2 − · · ·
n=1
n 2 3 4

TAN BAN PIN MA2311 72 / 86

Theorem (Alternating Series Test)

∞
X ∞
X
Let (−1)n+1 an (or (−1)n an ) be an alternating series.
n=1 n=1
Suppose that
1. an > 0 for all n.
2. {an } is decreasing (i.e. an > an+1 for all n), and
3. lim an = 0.
n→∞
∞
X ∞
X
n+1
Then (−1) an (resp. (−1)n an ) is convergent.
n=1 n=1

TAN BAN PIN MA2311 73 / 86

 ∞
X
Given a series an , if we take the absolute values of its terms, we obtain
n=1
∞
X
the series |an |.
n=1
∞
X (−1)n+1
Example: If the given series is , then taking the absolute
n
n=1
values of its terms gives the harmonic series:
∞ ∞
X (−1)n+1 X 1
= .
n=1
n n=1
n

TAN BAN PIN MA2311 74 / 86

Natural questions

∞
X ∞
X
1. Is it true that an converges ⇒ |an | converges?
n=1 n=1
X∞ X∞
2. Is it true that |an | converges ⇒ an converges?
n=1 n=1

TAN BAN PIN MA2311 75 / 86

Natural questions

∞
X ∞
X
1. Is it true that an converges ⇒ |an | converges?
n=1 n=1
X∞ X∞
2. Is it true that |an | converges ⇒ an converges?
n=1 n=1
∞
X (−1)n+1
Counter-example for (1): The series converges by the
n=1
n
∞ ∞
X (−1)n+1 X 1
alternating series test, but the series = diverges. So
n=1
n n=1
n

∞
X ∞
X
an converges ; |an | converges
n=1 n=1

TAN BAN PIN MA2311 76 / 86

Definition

∞
X
1. We say that the series an converges absolutely if the series
n=1
∞
X
|an | converges.
n=1
∞
X
2. We say the series an converges conditionally if
n=1
∞
X ∞
X
(i) an converges, and (ii) |an | diverges.
n=1 n=1

TAN BAN PIN MA2311 77 / 86

Examples on Absolute and conditional Convergence

∞
X (−1)n+1
1. The series converges absolutely, since
n=1
n2

∞ ∞
X (−1)n+1 X 1
=
n=1
n2 n=1
n2

is convergent (being the p-series with p = 2 > 1.)

∞
X (−1)n+1
2. We saw that the series converges, but
n=1
n
∞ ∞ ∞
X (−1)n+1 X 1 X (−1)n+1
= diverges. Hence, converges
n=1
n n=1
n n=1
n
conditionally.

TAN BAN PIN MA2311 78 / 86

 ∞
X
Theorem. If the series an converges absolutely, then it converges.
n=1
Remark: The above theorem suggests that one way to test a series for
∞
X
convergence is to first test it for absolute convergence. Note that |an |
n=1
is a non-negative series, thus one may potentially use the tests in previous
∞
X
sections to determine the convergence/divergence of |an |.
n=1

TAN BAN PIN MA2311 79 / 86

Example

From the previous example on Absolute and Conditional Convergence (1),

∞
X (−1)n+1
we saw that the series converges absolutely. Thus
n=1
n2
∞
X (−1)n+1
itself also converges.
n=1
n2

TAN BAN PIN MA2311 80 / 86

Theorem. Every series is either absolutely convergent, conditionally
convergent or divergent.

TAN BAN PIN MA2311 81 / 86

Theorem. Every series is either absolutely convergent, conditionally
convergent or divergent.
Example.
∞
X (−1)n
1. Determine whether the series √ is absolutely convergent,
n=1
n
conditionally convergent, or divergent.
∞ ∞
X (−1)n X 1
Sketch of Solution: √ = √ diverges (being the p-series
n=1
n n=1
n
1
with p = 2 < 1).
On the other hand, by the Alternating Series Test (check the three
∞
X (−1)n
conditions!), √ itself is convergent.
n=1
n
∞
X (−1)n
Thus √ converges conditionally.
n=1
n

TAN BAN PIN MA2311 82 / 86

Example.
∞
X cos n
2. Determine whether the series is absolutely convergent,
n=1
n3
conditionally convergent, or divergent. (Try!)

TAN BAN PIN MA2311 83 / 86

Some remarks on the various tests for convergence/divergence of
series
1. n-th term test:
a test for divergence ONLY, and no conclusion if lim an = 0, e.g.
n→∞
∞
X
(−1)n .
n=1
2. comparison Test (CT) / Limit Comparison Test (LCT):
when applying these tests, one usually compares the given series with
a geometric series or a p-series.
applicable to certain series which look like the geometric series or the
∞ ∞ 1
X 2(−1)n X 2n
p-series, e.g. , .
n=1
4n n=1
n2
when both the CT and LCT are applicable to a given series, it is
usually simpler and easier to apply the LCT.
∞
X 2 + (−1)n
when an oscillating factor/term appears, e.g. , try the
n=1
3n
CT rather than the LCT.
TAN BAN PIN MA2311 84 / 86
Some remarks on the various tests for convergence/divergence of
series
3. Ratio Test:
applicable to certain series which look like the geometric series, series
∞ 2 X
X n ∞ (2n)!
with n!, and certain series defined recursively, e.g. , .
n=1
3n n=1 4n · n!
∞
X  
1 1
an , where a1 = 1, an = 2 + n an−1 , n = 2, 3, · · · .
n=1
4. Root Test:
applicable to certain series where it involves a high power such as the
∞ ∞  2
n X 1 n
X 
n
n-th power, e.g. , 2 1− .
n=1
3n n=1 n

TAN BAN PIN MA2311 85 / 86

Some remarks on the various tests for convergence/divergence of
series
5. Alternating Series Test:
∞
X 1
applicable to certain series alternating series only, e.g. (−1)n √ .
n=1
n

TAN BAN PIN MA2311 86 / 86