Sequences and Series

By Nguyen Cao Manh Thang (Huynh Kim Gia Bao edited) from USTH Learning Support

December 2024

Contents
1 Sequences 2
1.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Define a sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 More on Divergence and Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.1 A formal definition - You can skip it! . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.2 Properties of limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.3 Subsequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Other properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Series 6
2.1 Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Important series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Series properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3 Series properties 8
3.1 First Divergence test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Geometric Series test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Study of non-negative series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3.1 Integral Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3.2 p-series test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3.3 Direct comparison test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3.4 Limit comparison test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.4 Ratio Test, Root Test and Absolute Convergence . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4.1 Ratio Test and Root Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4.2 Absolute Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.5 Alternating Series Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.6 Convergence Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Power Series. Taylor and Maclaurin Series 15

4.1 Power Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2 Interval of Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3 Series Manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.1 Subtitution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.2 Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3.3 Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.4 Taylor and Maclaurin Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.5 Applications of Taylor Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.5.1 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.5.2 Error Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

1
1 Sequences
1.1 Definitions
Sequences are things from which series are built. A sequence is nothing but an infinite list of numbers,
written in a definite order:
a1 , a2 , a3 , a4 , ...
To say a sequence converges to a number L is to say that the terms of the sequence get closer and closer to
L the further along the sequence we go. Written in terms of limits, we have:

lim an = L
n→∞

A sequence diverges just means that it does not converge. Let’s take a look at some examples:
1. Consider the sequence:
(−1)n
an = 2 +
n2
The first few terms of this sequence are:
−1 1 −1 1
2+ , 2 + 2 , 2 + 2 , 2 + 2 , ...
12 2 3 4
n
As n goes to infinity, the (−1)
n2 gets closer and closer to zero, since the numerator bounces back forth
between −1 and 1 while the denominator gets larger and larger. Thus:

(−1)n
 
lim an = lim 2 + =2+0=2
n→∞ n→∞ n2

We conclude this sequence converges to 2.

2. Consider the sequence:
2 · 3n+1 (3n − 1)!
an = n
+
5 (3n + 1)!
We can rewrite the first part of the sum as:
 n
2 · 3n+1 2 · 3 · 3n
3
= =6
5n 5n 5

n
Observe that this sum converges to 0, as when n gets larger, 35 gets smaller and smaller. Thus:
 n
2 · 3n+1 3
lim = lim 6 =0
n→∞ 5n n→∞ 5

Now, the second part of the sum in the definition of an can be written as:

(3n − 1)! (3n − 1)! 1

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

It is clear that:
(3n − 1)! 1
lim = lim =0
n→∞ (3n + 1)! n→∞ (3n + 1)(3n)
Since both parts of the sum in the definition of an individually converge, we can say that:

2 · 3n+1 2 · 3n+1
 
(3n − 1)! (3n − 1)!
lim an = lim + = lim + lim =0+0=0
n→∞ n→∞ 5n (3n + 1)! n→∞ 5n n→∞ (3n + 1)!

so, our given sequence an converges to 0.

2
1.2 Define a sequence
Some different ways to define a sequence include:
1. Sequences can be defined by giving a formula for the nth term.
∞
(−1)n n=1 = −1, 1, −1, 1 : an = (−1)n

2. Sequences may be defined by different expressions for different subsets of indices (like piece-wise func-
tions). (
1 1 ≤ n ≤ 100
cn , with cn = √
ln( n − 10) n ≥ 101

3. Sequences may be defined recursively: The Fibonacci sequences F1 = F2 = 1, Fn = Fn−1 + Fn−2

4. Some sequences do not have a simple defining expressions. Example: The sequence pn , where pn is
the n-th prime number.
Why learn this? One type of question that may appear in the test is: Determine whether a sequence
diverges/converges, but a sequence is not given in the form of the nth term but in the recursive method,
for example. It is important that you can try to work that out. Furthermore, we will soon define the
convergence property of series based on that of sequence, so you should understand, to the least, what does
”the convergence of a sequence” actually mean!

Exercise 1: Can you define the sequence of all positive even integers using the four aforementioned methods?
√
Exercise 2: Define a sequence x1 , x2 , x3 , x4 ... by setting x1 = 2 and then recursively defining each other
term in terms of the previous one via:
√
xn+1 = 2 + xn for n ≥ 1

Given that this sequence does diverge, can you find where does it converge to?

1.3 More on Divergence and Convergence

1.3.1 A formal definition - You can skip it!
(Feel comfortable to skip this - I only include for those who are ”curious”!)

A sequence is convergent if and only if for every ϵ > 0, there exists a number N so that for every m, n ≥ N ,
|am − an | < ϵ.

Remark: This is quite similar to the notion of limits. In most textbooks for pure Mathematics guys,
Sequences are often discussed before the notion of Limits, Differentiations and Integration. You can verify it
by checking ”Principles of Mathematical Analysis” by Walter Rudin (famously known as the ”Baby Rudin”
book - it’s not easy as the ”baby” title may sound :)).

1.3.2 Properties of limits

As we can define a sequence with the idea of a limit, it is possible that the properties of limits also apply to
the properties of a sequence. Let’s review this:

3
For the other limit properties, please review Chapter 1 - Section ”Limits”.

1.3.3 Subsequences
A subsequence of {an } is a sequence ”extracted” from the sequence {an } For examples:
• {an2 }, that is a1 , a4 , a9 , a16 , ...
• The even subsequence {a2n } that is a2 , a4 , a6 , ..., a2n ... The odd subsequence {a2n−1 } that is a1 , a3 , a5 , ..., a2n−1 ...
A useful property of subsequences is this theorem (Bolzano - Weirstrass Theorem):

If lim an = L, then every subsequences of {an } also converges to L.

n→∞

4
1.4 Other properties
Bounded sequence. A sequence an is said to be bounded if there are numbers M and m such that:

m ≤ an ≤ M, for every n.

Numbers m and M are called a lower bound and an upper bound of the sequence an .
If a sequence an converges, then it is bounded!

Monotone sequences
• A sequence an is said to be increasing for n ≥ n0 if an ≤ an+1 for every n ≥ n0 (where n0 is some
integer).
• A sequence an is said to decrease for n ≥ n0 if an ≥ an+1 for every n ≥ n0 (where n0 is some integer).
• A sequence an is said to be monotonic if it is increasing or decreasing.
Every bounded, monotonic sequence is convergent.

5
2 Series
2.1 Definition
Let {an } be a sequence. An infinite series is a formal expression (of an “infinite sum”) of the form:
∞
X
= a1 + a2 + a3 + ...
n=1

To be clear, what makes this as “infinite” sum is the fact that we are adding infinitely many things, NOT
that the resulting sum itself could be infinite. The series does not have to start at 0; for instance:
∞
X
= a2 + a3 + a4 + ...
n=2

begins the infinite sum at n = 2. In general, the notion says we plug in the first value of n to get the first
term, then increase n by 1 to compute the next term, then increase again and so on, adding on each new
term we get at each step.
The key question we care about is whether a series converges, meaning that we actually get a finite
value out of the given infinite sum, or diverges, meaning that we don’t get a specific value.

Partial Sums. But to make all of this precise, we have to be more careful about what it actually means
for a series to converge.
P∞
Consider the partial sums sk of the series n=0 in question, which are the sums obtained by adding one
more term in our series at each step:

s0 = a0
s1 = a0 + a1
s2 = a0 + a1 + a2
..
.
k
X
sk = a0 + a1 + a2 + a3 + ... + ak = an
n=0

The key point here is that if a series DOES have a finite sum, then as k tends to infinity, the partial sum
gets better at approximating the finite value of the series. We arrive at this definition:
P∞
The series n=1 is said to converge, if the sequenceP{sk } converges. Moreover, if limk→∞ Sk = S, then
∞
we say the given series converges to S, and write n=1 = S. The number S is called the sum of the
series. If a series does not converge, then it is said to diverge.

Example 1. We now investigate whether a series converges or not using the definition:
∞
X 1
n=1
n(n + 1)

This series is equivalent to:

∞  
X 1 1
−
n=1
n n+1

We can determine the k th partial sum by:

       
1 1 1 1 1 1 1 1
Sk = 1 − + − + − + ... + − =1−
2 2 3 3 4 k k+1 k+1
It is obvious that:  
1
lim Sk = lim 1− =1
k→∞ k→∞ k+1
thus, we conclude that the given series converges and its sum is 1.

6
2.2 Important series
Let’s have a look at some of the most important series that you have to remember in order to ace the Final
Test:
1. Geometric Series:
∞
X
arn = a + ar + ar2 + ar3 + ... + arn + ...
n=0

where a ̸= 0 is a constant number, and for any n >= 0, aan+1

n
= r is also a constant number. (known
as the common ratio r)

Practice 1: Prove that this series converge and find its sum:
∞
X 2 · 3n+1
n=0
5n

Practice 2: Check whether these series converge or diverge and explain why. If a series converge, find
its sum?

2a.
∞
X 4n+2
(−1)n
n=1
3n−1

2b.
∞  
X 1 1
+ n
n=2
2n 3

2. Harmonic series. Remember that this series diverges:

3. Alternating Series.

7
 4. p-series. Remember that p-series has this form:
∞
X 1
n p
n=1

where p is some constant. It converges when p > 1, and diverges if p ≤ 1. We will return to this in
the convergence/divergence test part.

2.3 Series properties

Finally, take a look at some basic series properties.

3 Series properties
As I have mentioned before, the key question regarding series is whether a series converges, meaning that
we actually get a finite value out of the given infinite sum, or diverges, meaning that we don’t get a
specific value.
Part 3 would deal with all important types of tests for divergence and convergence. We will also share a
strategy for testing this property of a series at the end of this section. This is indeed very important, so
please take a close look on this part!

3.1 First Divergence test

∞
X ∞
X
In a series an , if the sequence an does not converge to 0, the series an must diverge.
n=1 n=1

8
Example: Consider the series:
∞
X
n1/n
n=1

First and foremost, we should use the n-th term test (Or, in this document, First Divergence Test):

lim n1/n
n→∞

The limit leads to the indeterminate form (∞)0 . We let f (n) = n1/n and find limn→∞ ln f (n). Since:

ln n
ln f (n) = ln n1/n =
n
Applying the l’Hôpital rule gives:

ln n 1/n
lim ln f (n) = lim = lim =0
n→∞ n→∞ n n→∞ 1

Therefore,
lim n1/n = lim f (n) = lim eln f (n) = e0 = 1
n→∞ n→∞ n→∞

As the n-th term test shows that the sequence converges to 1, which is different from 0, we conclude that
the series diverges.

3.2 Geometric Series test

Recall that a geometric series has the form:
∞
X
arn = a + ar + ar2 + ar3 + ... + arn + ...
n=0

where a ̸= 0 is a constant number, and for any n ≥ 0, aan+1

n
= r is also a constant number. (known as the
common ratio r).
It is easy to determine whether a geometric series converges or diverges: return to the geometric series
introduction in section 2.2 to check this.

3.3 Study of non-negative series

These are series with non-negative terms:
∞
X
an , an ≥ 0 for all n.
n=1

There are special techniques for testing the convergence of these series.

9
3.3.1 Integral Test
The integral test applies to series of the form:
∞
X
f (n)
n=1

where f is a continuous, positive,

P∞ decreasing function R∞
In this setting, the series n=1 f (n) and the improper integral 1 f (x)dx both behave in the same way,
meaning they both converge or they both diverge. Therefore, we can turn problems about series convergence
into ones about integral convergence instead.

IMPORTANT. Remember to state that f (n) satisfies all the condition before doing the integral test
so that you would not lose 0.25 - 0.5pt in the test!

Example 1. Consider the harmonic series:

∞
X 1
n=1
n
This function satisfies the three conditions to apply the integral test: Continuous, Positive, and Decreasing.
This means that we can now apply the integral test. Consider the integral:
Z ∞
1
dx
1 x

We get:
Z ∞ Z b b
1 1
dx = lim dx = lim ln x = lim (ln b − 1) = ∞.
1 x b→∞ 1 x b→∞ 1 b→∞

We conclude this series diverges.

Example 2. Consider the series:

∞
X 3
n2 e−n
n=2

Observe:
• This series starts with n = 2. We should consider the integral from 2, not 1 like the previous example.
• We can prove that we can use the integral test, as this satisfies the three necessary conditions (Prove
it!)
Now we start doing the integral test:
Z ∞ Z b  
b
3 3 1 3
x2 e−x dx = lim x2 e−x dx = lim − e−x
2 b→∞ 2 b→∞ 3 2
 
1  −b3
= lim − e − e−8
b→∞ 3
1 −8
=− e .
3

This integral converges, so the integral test says that the given series also converges!

Using the integral test, determine whether these series converge or diverge
1.
∞
X n2
n=3
n3+1

2.
∞
X 1
n=2
n ln n

10
3.3.2 p-series test.
Remember that p-series has this form:
∞
X 1
n p
n=1

where p is some constant. It converges when p > 1, and diverges if p ≤ 1.

3.3.3 Direct comparison test.

The key facts to remember are:
• If the larger series converges, so does the smaller one.
• If the smaller series diverges, so does the larger one.
• We must ”guess” the proper series to do this - therefore, this is a hard method.
Example 1. Consider the series:
∞
X 10n2 − 3n − 1
n=1
n4 + n2 + 1
Observe that:
∞
X 10n2 − 3n − 1 10n2 10
0≤ 4 2
≤ 4
= 2
n=1
n +n +1 n n
Sine the sum on the right and the left are all finite, the infinite sum should be finite. We conclude the given
series converges by the direct comparison test.
Example 2. Consider the series:
∞
X cos2 (n)
n=1
n3/2
Observe that:
∞ ∞
X cos2 (n) X 1
≤ (As 0 ≤ cos2 (n) ≤ 1)
n=1
n3/2 n=1
n3/2
P∞ 1
P∞ cos2 (n)
By the p-series test, n=1 n3/2 converges. Therefore: n=1 n3/2
converges

3.3.4 Limit comparison test.

P P
This test only applies to series consisting of positive terms. For such series an and bn , look at the limit:
an
L = lim
n→∞ bn
P P
If L exists and is positive, series an and bn behaves in the same way: Both will converge or diverge.
Example. Consider the series:
∞
X n(1 − e−n )
n=5
n3 + 3
Focusing on dominant terms only suggests that this series should behave like:
∞ ∞
X n X 1
3
= 2
n=5
n n=5
n

We use the limit comparison test. Look at the limit:

n(1−e−n )
n3 +3
lim 1
n→∞
n2
P 1
We can prove that the limit is 1. As the limit is finite and positive, since n2 converges, our given series
does converge!
Warning. When limit L = 0, we can’t determine whether the two series behave in the same way or not.

11
3.4 Ratio Test, Root Test and Absolute Convergence
3.4.1 Ratio Test and Root Test

We take a look at some examples:

Example 1. Consider the series:

∞
X n
(−1)n
n=1
5n

In the notation of the ratio test, an = (−1)n 5nn , so we compute the limit:

|an+1 | (−1)n+1 5n+1

n+1 (n + 1)5n n+1 1
lim = lim n = lim n+1
= lim = .
n→∞ |an | n→∞ n
(−1) 5n n→∞ n5 n→∞ 5n 5

Since this limit is less than 1, the ratio test tells us that this series converges

Example 2. Consider the series:

∞
X (−3)n
n=0
(2n + 1)!
We compute the following limit:
(−3)n+1
(2(n+1)+1)! 3n+1 (2n + 1)! 3
lim (−3)n
= lim n
= lim = 0.
n→∞ n→∞ (2n + 3)! 3 n→∞ (2n + 3)(2n + 2)
(2n+1)!

Since we got a limit of zero, the ratio test implies that our given series converges.

Example 3. Consider the series:

∞   n2
X 1
(−1)n 1 −
n=1
n
We do the root test: v
u   n2  n
u
n n
1 1
lim t (−1) 1 − = lim 1 −
n→∞ n n→∞ n

Note that ex = (1 + nx )n . Therefore:

 n
1 1
lim 1 − = <1
n→∞ n e
According to the root test, this series converges absolutely.

12
3.4.2 Absolute Convergence

Why investigate absolute and conditional convergence? For the sake of brevity, I will not include
this in this document. However, you can find an explanation in the appendix of this document, or read the
books if you are curious enough!

13
3.5 Alternating Series Tests

IMPORTANT.Remember to state that bn+1 ≤ bn in your exam so that you would not lose 0.25pt in the
test.

3.6 Convergence Strategies

• Use first and foremost the n-th term test.
• If the given series has the form of geometric or p-series, use the corresponding test for them. This
applies to alternating series test too.
• If not possible, try ratio/root test next. This is highly effective for series with factorials!
• Try limit comparison test next if nothing has worked so far. This works especially with fractions whose
numerator and denominators only involve powers of n.
• Integral test may be used, but remember to check all the necessary conditions: positive, continuous
and decreasing function.
• Direct comparison test might be the toughest one. However, for any series with involves sine or cosine,
perhaps this is the first test you should use, as sine and cosine terms are usually simple to bound.
Exercise: Determine whether these series converges or diverges?
∞ ∞ ∞
X 2n X 1 X 1
1. 2. 3. √ √
n=1
3n n=1
2
n + 30 n=1
2 n+ 3n
∞ ∞ ∞
X sin2 n X 2n X 1
4. 5. 6.
n=1
2n n=1
3n − 1 n=3
ln (ln n)
∞ ∞ ∞ √
X 3n X 3n + 1 n
 X n3 + n
7. 8. 9.
n=1
n! n=1
2n n=1
n4 − n2

14
4 Power Series. Taylor and Maclaurin Series
4.1 Power Series
∞
1 X
= xn where |x| < 1
1 − x n=0
Now, we view the right side as a function by treating x as a variable:
P∞ plugging in a value of x into 1this series
gives the value on the left side. We can say that the series n=0 xn represents the function 1−x on the
interval (-1, 1). Representing functions as series is the whole reason why we care about series in the first case!

Example 1. The power series:

∞
X
(3x − 1)n
n=0
As we have:
∞
1 X
= yn
1 − y n=0
Let y = 3x − 1:
∞
1 X
= (3x − 1)n
1 − (3x − 1) n=0
This series will converge when |y| < 1, so when:
1 1 2
|3x − 1| < 1 ⇔ x − < ⇔0<x<
3 3 3
We conclude that the given series converges for x in 0, 23 , thus:

∞ ∞  n
1 X X 1
= (3x − 1)n = 3n x −
2 − 3x n=0 n=0
3
1 1
on the interval 0, 32


represents the function 2−3x as a power series centered at 3

4.2 Interval of Convergence

We test a power series for convergence by using various tests to see where it converges and diverges. For
the geometric series, for example, it diverges when |r| > 1, and converges when |r| < 1. How about for other
series?
One key fact is that any power series converges for x in some interval around its center a. There are three
possible cases:

15
 • Converges absolutely for all x
• Converges at only one point x = a
• Converges when |x − a| < R and diverge when |x − a| > R, with R is a positive real number.
We now get to some definitions:

Find the interval and radius of convergence is an application of the ratio test! It is on this interval that we
can make sense of saying that the power series in question represents a function, since this interval charac-
terizes the values of x we can actually plug into that function and have the resulting value make sense.

Example 1. Find the interval of convergence of this series:

∞
X
n(x − 2)n
n=1

Using the ratio test:

(n + 1)|x − 2|n+1
 
|an+1 | (n + 1)|x − 2| 1
lim = lim = lim = |x − 2| lim 1 + = |x − 2|.
n→∞ |an | n→∞ n|x − 2|n n→∞ n n→∞ n

According to the ratio test, the given series converges when |x − 2| < 1, and diverges when |x − 2| > 1. In
other words, the series converges at least for x ∈ (1, 3).
Recall that the ratio test is inconclusive when limn→∞ |a|an+1
n|
|
= 1. We must test the case when x =
1 and when x = 3. Testing gives us that when x = 1 or x = 3, our series diverges. Only after this step
can we conclude the interval of convergence is (1,3).

Example 2. Find the interval of convergence of this series:

∞
X
(2n + 1)!(x − 1)n
n=0

16
Using the ratio test:

(2(n + 1) + 1)!|x − 1|n+1 (2n + 3)!

lim n
= lim |x − 1| = lim (2n + 3)(2n + 2)|x − 1|
n→∞ (2n + 1)!|x − 1| n→∞ (2n + 1)! n→∞

We can see that (2n + 3)(2n + 2) goes to infinity for any x except x = 1. This is less than 1, so the ratio
test tells us that the given series does converge when x = 1.
Conclusion: The series converges only when x = 1, and thus the interval of convergence consists of just the
single point 1.

Example 3. Find the interval of convergence of this series:

∞
X xn
n=0
n!

Using the ratio test:

|x|n+1 |x|
lim = lim =0
n→∞ (n + 1)!|x|n n→∞ n + 1

This is smaller than 1 no matter what x is, we conclude that series:

∞
X xn
n=0
n!

does converges for all x. The interval of convergence is thus (−∞, ∞), and we say that the radius of
convergence is ∞. In this case there are no endpoints to check!

4.3 Series Manipulation

4.3.1 Subtitution
From the basic power series:
∞
1 X
= yn
1 − y n=0

we can substitute y = f (x) on both sides. This is what we call Substitution.

Example 1. What function does this series represent?

∞
X
(−1)n x2n
n=0

From the basic power series, setting y = −x2 gives:

∞
X ∞
X
(−x2 )n = (−1)n x2n
n=0 n=0

which is precisely the series we’re asking about. We thus get:

1 1
= = sum∞ n 2n
n=0 (−1) x
1 − (−x2 ) 1 + x2

This is valid on the interval (-1, 1).

4.3.2 Differentiation
You can differentiate both sides of the basic power series to get some new series. This process is called
term-by-term differentiation, and is represented by:

17
Example 1. Differentiate the standard geometric series:
∞
1 X
= xn
1 − x n=0

Solution:
∞
  !
d 1 d X
n
= x
dx 1−x dx n=0
∞
1 X d n
= x
(1 − x)2 n=0
dx
∞
X
= nxn−1
n=1
∞
X 1
= nxn
n=1
x

To be clear, we start the final series at n = 1 since the n = 0 term is zero anyway, so it is not worth writing.
Multiplying through by x would then give the following series representation:
∞
x X
= nxn
(1 − x)2 n=1

Example 2. Consider the series:

∞
X
(−1)n 2nx2n−1
n=1

The key observation here is that:

∞
!′ ∞
X X
n 2n
(−1) x = (−1)n 2nx2n−1
n=0 n=1

Since:
∞
1 X
= (−1)n x2n
1 + x2 n=0

the series we’re looking at in this example should represent the function obtained by differentiating the
left-hand side. By doing this:
∞
−2x X
= (−1)n 2nx2n−1
(1 + x2 )2 n=1
∞
x X
or − = (−1)n nx2n−1
(1 + x2 )2 n=1

18
Exercise: Which function is represented by this series:
∞
X
n(n − 1)xn
n=2

4.3.3 Integration
Similarly, integrating both sides of the series is valid, and a summary for this process is here:

You can take a look at this example:

19
4.4 Taylor and Maclaurin Series
Definition. If we want to express f as a power series centered at a:
∞
X
f (x) = cn (x − a)n ,
n=0

f (n) (a)
the coefficients needed must be given by cn = n! . The resulting power series:
∞
X f (n) (a)
(x − a)n
n=0
n!

is called the Taylor series of f centered at a and is the only power series which could possibly equal the
function f (x) on its interval of convergence. Taylor series which are centered at 0 will be called Maclaurin
series
∞
X f (n) (0) n
x
n=0
n!

There are several important things you should learn by heart:

• In general, the Taylor series generated by a function may or may not be f (x).
• However, the theorem (about power series representation) says that if f is representable as a power
series at a, then the power series is ı́t Taylor series.
• Many functions have power series representation, e.g, 1
1+x , ln (1 + x), ex , sinx, (1 + x)α
• However, there exist functions that are differentiable but has no power series representation!
Taylor polynomials. The Taylor polynomial of degree n of f , denoted by Tn (x), is defined as follows:

f ′ (a) f ′′ (a) f (k) (a) f (n) (a)

Tn (x) = f (a) + (x − a) + (x − a)2 + ... + (x − a)k + .... + (x − a)n
1! 2! k! n!

The coefficient of (x − a)k is:

f (k) (a)
ck = (k = 0, 1, 2, ..., n)
k!

20
Example 1. Find the Taylor series and the Taylor polynomials generated by f (x) = ex at x = 0

f (n) (x) = ex and f (n) (0) = 1 for every n = 0, 1, 2...

Therefore, the Taylor series generated by f at x = 0 is:

f ′′ (0) 2 f (n) (0) n

f (0) + f ′ (0)x + x + ... + x + ....
2! n!
x2 xn
=1+x+ + ... +
2 n
∞
X xk
=
k!
k=0

The Taylor polynomial of order n at x = 0 is:

x2 xn
Pn (x) = 1 + x + + ... +
2 n!

Example 2. Find the Taylor series and Taylor polynomials generated by f (x) = cosx at x = 0
Solution. The cosine and its derivatives are:

f (x) = cosx f ′ (x) = −sinx

f ′′ (x) = −cosx f (3) (x) = sinx
.. ..
. .
f (2n) (x) = (−1)n cosx f (2n+1) (x) = (−1)n+1 sinx

The Taylor series generated by f at 0 is:

f ′′ (0) 2 f (n) (0) n

f (0) + f ′ (0)x + x + ... + x + ...
2! n!
x2 x4 x2n
= 1 + 0.x − + 0 · x3 + + ... + (−1)n + ...
2! 4! (2n)!
∞
X (−1)k x2k
=
(2k)!
k=0

Taylor remainder. Suppose a function f has derivatives of all orders in an open interval l containing a.
Then for each x ∈ I and each positive integer n, we have:

f (n+1) (c)
f (x) − Tn (x) = (x − a)n+1
(n + 1)!

where c is some value between a and x.

The expression:
f (n+1) (c)
Rn = (x − a)n+1
(n + 1)!
is called the Taylor remainder of order n or the error term for the approximation of f by Tn (x).

4.5 Applications of Taylor Series

This section contains difficult exercise - According to Prof. Khoi Le Hai, these exercises are very difficult,
and should you see them appearing in your exam, consider leaving them until you have done the other
problems. Therefore, if you aim to just ”pass” the subject, I highly recommend not reading this section and
revising the previous parts of this sequences - series.

21
4.5.1 Convergence

In other word, given a function, you first write the Taylor remainder of the Taylor series generated by this
function. After that, you should determine the limit of this remainder.

Example. Show that the Taylor series for cos x at x = 0 converges to cos x for every value of x.

Solution. We add the remainder term to the Taylor polynomial for cos x to obtain Taylor’s formula
for cos x with n = 2k:
x2 x4 x2k
cos x = 1 − + − ... + (−1)k + R2k (x)
2! 4! (2k)!
Because the derivatives of the cosine have absolute value less than or equal to 1, the Remainder Estimation
Theorem with M = 1 gives:
|x|2k+1
|R2k (x)| ≤ 1 ·
(2k + 1)!
For every value of x, R2k → 0 as k → ∞. Therefore, the series converges to cos x for every value of x. Thus,
∞
X (−1)k x2k x2 x4 x6
cos x = =1− + − + ...
(2k)! 2! 4! 6!
k=0

4.5.2 Error Estimation

22
Example. Estimate the value if P4 (x) = 1 + x + (x2 /2) + (x3 /6) + (x4 /24) is used to estimate the value
of ex at x = 1/2.

Solution Recognize that P4 (x) is the Taylor polynomials of ex at degree 4. Recall that when we approximate a
function f (x) by its Taylor polynomial Tn (x), the Taylor’s remainder Rn provides the error in the estimation.

Therefore, the error if P4 (x) is used to estimate ex at x = 1/2 is:

f (n+1) (c)
R4 (x) = (x − a)n+1 for n = 4 in our exercise.
(n + 1)!
f (5) (c)
= (0.5)4+1
5!
ec
= (0.5)5
5!
For c ∈ [0, 0.5], ec ≤ e0.5 or ec ≤ 2.7 The error is:

R4 ≤ 7.03 · 10−4

Extra Notes. You may take a look at the Alternating Series Estimation Theorem to deal with these
exercises:

23