Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

CHAPTER 4: SEQUENCES AND SERIES

4.1 Sequences

4.1.1 Convergence and divergence of Sequences

A. Definition of Sequence

Literally, a sequence is an ordered list of objects. The number of objects in a sequence can be finite or infinite.
An infinite sequence is an ordered and unending list of objects. More formally

Definition 1 [sequence]: Let X be any set. An infinite sequence (or just a sequence) in X is a function whose
domain is the set of all integers greater than or equal to a given integer m.

In short, a sequence is a function of the form:

𝑓: {𝑚, 𝑚 + 1, 𝑚 + 2, ⋯ } → 𝑋

Remark 1 (range of a sequence): The range of a sequence can be a subset of any nonempty set X. The
sequence whose range is a subset of ℝ is called a real sequence. Thus a real sequence is a function 𝑓 of the
form:

𝑓: {𝑚, 𝑚 + 1, 𝑚 + 2, ⋯ } → ℝ

All sequences considered in this chapter are real sequences.

Remark 2 (notation of a sequence and its values): If 𝑓 is a sequence and 𝑛 is in the domain of 𝑓, then we
write 𝑓𝑛 instead of 𝑓(𝑛) . Moreover f itself is denoted by {fn } ∞
n=m or just by {fn } . We can use other letters

(specially 𝑎, 𝑏, 𝑐 ) instead of 𝑓.

Remark 3: Let {an } ∞

n=m be a sequence. Then,

- the numbers 𝑎𝑚 , 𝑎𝑚+1 , 𝑎𝑚+2 ,. ... are called the terms,

- 𝑛 is called the running index and
- 𝑚 is called the initial index of the sequence.

Example 1:

1
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

1
(a) The function 𝑓(𝑛) = , 𝑛 ∈ ℕ, is a sequence. It establishes the following ordered list.
𝑛

𝑓 (1) = 1 (first object)

𝑓 (2) = ½ (second object)

1
𝑓 (3) = 3
(third object)

𝑛−1
(b) The function 𝑎(𝑛) = , 𝑛 ∈ {−5, −4, −3, ⋯ }, is a sequence.
𝑛2 +1

B. Descriptions of Sequences

A sequence {𝑎𝑛 } can be described by:

a) listing only the first few terms and by putting three dots after them i.e., in the form 𝑎𝑚 , 𝑎𝑚+1 , ⋯

b) giving a formula for 𝑎𝑛 terms of 𝑛, i.e., in the form 𝑎𝑛 = 𝑓(𝑛)

c) using concise notations of the form {an } ∞

n=m or just {a n } where 𝑎𝑛 is given by a formula involving n.

d) Verbal methods, i.e., by using ordinary English words.

e) a recursive formula, i.e., specifying the first few terms of the sequence explicitly and expressing each
term after them by a formula involving its predecessors.

f) Pictorial (graphical) methods i.e., by plotting (𝑛, 𝑎𝑛 ) on 𝑥𝑦- plane or by plotting 𝑎𝑛 on a number line
for the first few values of n.

Example 2:

1
(I) Describe the sequence 𝑎: ℕ → ℝ given by 𝑎(𝑛) = 𝑛 by using extended list, explicit formula, concise

notation and verbal methods.

Answers:

1 1
a) Extended list : 1, 2 , 3 , …

2
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

1
b) Explicit formula: 𝑎𝑛 = , 𝑛 ∈ ℕ
𝑛

1
c) Concise notation: {n } ∞
n=1

d) Verbal method: the sequence of sequence of reciprocals natural numbers.

The above sequence is called harmonic sequence.

(II) Recursive formula: Let f1 = 0, f2 = 1, fn+1 = fn−1 + fn for 𝑛 ≥ 2. This sequence is called
the Fibonacci sequence
𝑛+1 ∞
(III) Graph: The next graph gives the graph of the sequence { } for the first 30 terms.
𝑛2 +1 𝑛=1

𝒏+𝟏 ∞
Figure 1: The graph of {𝒏𝟐 +𝟏} for the first 30 terms
𝒏=𝟏

C. Definition of Convergent and Divergent Sequences

Definition 2 [limit of a sequence]: Let {an }∞ ∞

n=m be a sequence. A number 𝐿 is said to be the limit of {a n }n=m if

∀𝜀 > 0, ∃ a positive integer N such that

|𝑎𝑛 − 𝐿| < 𝜀, ∀ 𝑛 ≥ 𝑁.

3
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

If such L exists, we say {an }∞

n=m converges (or is convergent) to L. Otherwise, we say it diverges (is divergent).

If {an }∞
n=m converges to L, we write lim𝑛→∞ 𝑎𝑛 = 𝐿 or 𝑎𝑛 → 𝐿.

Remark 1: Some authors use the inequality "𝑛 > 𝑁" rather than "𝑛 ≥ 𝑁" in the above definition. The two
definitions are equivalent.

Remark 2: If lim𝑛→∞ 𝑎𝑛 = 𝐿 , then every neighborhood (𝑎 − 𝜀, 𝑎 + 𝜀) of L contains all but finitely ,many
𝑥𝑛 ’𝑠.

4.1.2 Properties of Convergent Sequences

Theorem 1 [Uniqueness of limit of a sequence]: If a sequence converges, the limit is unique.

n
Example 3 (convergent sequence): Use the definition, to show that {3n2 +1}∞
n=1 converges to 0.

Theorem 2: Let {an }∞ ∞

n=m and Let {bn }n=m be convergent sequences. If 𝑎𝑛 < 𝑏𝑛 ∀𝑛, then

lim𝑛→∞ 𝑎𝑛 ≤ lim𝑛→∞ 𝑎𝑛 .

Caution: As given in Theorem 2, limit of sequence does not preserve strict inequality.

Example 4 (convergent sequence): Let 𝑎𝑛 = 𝑐, 𝑛 ∈ ℕ, where c is a constant . Use the definition, to show that
𝑎𝑛 → 𝑐.

(−1)n ∞
Example 5 (convergent sequence): Use the definition, to show that { }n=1 converges to 0.
n

Definition 3 [divergence of a sequence]:: Let {an }∞ ∞

n=m be a sequence. We say{a n }n=m

a) diverges to infinity( ∞) if for every real number M there exists a positive integer N such that

4
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

𝑎𝑛 > 𝑀, ∀ 𝑛 ≥ 𝑁.

b)diverges to negative infinity ( −∞)if for every real number M there exists a positive integer N such that
𝑎𝑛 < 𝑀, ∀ 𝑛 ≥ 𝑁.

c) oscillates (is oscillating) if it diverges but does not diverge to ∞ or −∞.

Example 6: Use the definition, to show that {n2 }∞

n=1 diverges to ∞ .

Example 7(oscillating sequence): Show that {(−1)n }∞

n=1 diverges and it does not diverge to ±∞.

Theorem 3[convergence of combinations of sequences]: Let {an }∞ ∞

n=m and Let {bn }n=m be

convergent sequences. Then the sequences {an + bn }∞ ∞ ∞

n=m , {a n − bn }n=m and {a n bn }n=m are
a
convergent. If lim𝑛→∞ 𝑏𝑛 ≠ 0, then {𝑏n }∞
n=m is also convergent. Moreover:
𝑛

(i) lim𝑛→∞ [𝑎𝑛 + 𝑏𝑛 ] = lim𝑛→∞ 𝑎𝑛 + lim𝑛→∞ 𝑏𝑛

(ii) lim𝑛→∞ [𝑎𝑛 − 𝑏𝑛 ] = lim𝑛→∞ 𝑎𝑛 − lim𝑛→∞ 𝑏𝑛

(iii) lim𝑛→∞ [𝑎𝑛 𝑏𝑛 ] = [ lim𝑛→∞ 𝑎𝑛 ][ lim𝑛→∞ 𝑏𝑛 ]

a lim 𝑎
(iv) lim𝑛→∞ 𝑏n = lim𝑛→∞ 𝑏𝑛 (provided that lim𝑛→∞ 𝑏𝑛 ≠ 0, )
𝑛 𝑛→∞ 𝑛

Proof: Follows from the proof of the corresponding limit theorems for functions.

Corollary: If {an }∞ ∞
n=m is convergent and K is a constant, the sequence {Ka n }n=m is convergent and

lim𝑛→∞ 𝐾𝑎𝑛 = 𝐾 lim𝑛→∞ 𝑎𝑛

Proof: Put 𝑏𝑛 = 𝐾 and apply Thorem 2(iii) and Example 4. ∎

Computing limits of sequences fast by using dominant term analysis:

When finding the limit of a function at infinity, it is only necessary to consider the dominant term. A term 𝑓𝑘 (𝑥)
𝑓 (𝑥)
of 𝑓(𝑥) = 𝑓1 (𝑥) + ⋯ + 𝑓𝑛 (𝑥) is said to be dominant term if lim𝑥→∞ 𝑓 𝑗 (𝑥) = 0 for 𝑗 = 1, … , 𝑛 and 𝑗 ≠k.
𝑘

When treating quotients of functions, we only need to consider the dominant terms in the numerator and
𝑎 𝑑𝑜𝑚𝑖𝑛𝑎𝑡𝑖𝑛𝑔 𝑡𝑒𝑟𝑚 𝑜𝑓 𝑎
denominator. lim𝑛→∞ 𝑎𝑛 = lim𝑛→∞ [𝑑𝑜𝑚𝑖𝑛𝑎𝑡𝑖𝑛𝑔 𝑡𝑒𝑟𝑚 𝑜𝑓 𝑏𝑛 ] and lim𝑛→∞ 𝑏𝑛 = lim𝑛→∞ 𝑑𝑜𝑚𝑖𝑛𝑎𝑡𝑖𝑛𝑔 𝑡𝑒𝑟𝑚 𝑜𝑓𝑏 𝑛
𝑛 𝑛

5
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Example 8[Computing limits of sequences using dominant term analysis and quotient rule]:

𝑛3 +4𝑛2 −1
Does the sequence 𝑎𝑛 = converge or diverge?
2−4𝑛4

Solution: The dominant (highest degree) term in the numerator is 𝑛3 . The dominant (highest degree) term in the
denominator is −4𝑛4 . We can factor out the largest terms from both the numerator and denominator and do a
little algebra.

4 1
𝑛3 + 4𝑛2 − 1 𝑛3 [1 + 𝑛 − 3 ]
= 𝑛
2 − 4𝑛 4 1
−4𝑛4 [ + 1]
−2𝑛4

The second term becomes arbitrarily close to 1 as 𝑛 grows larger and larger, so the limit of the sequence is
completely determined by the ratio of the highest degree term in the numerator to the highest degree term in the
1 𝑛3
denominator. In this case, that ratio is − 4𝑛 and it converges to 0. Therefore, lim𝑛→∞ 𝑎𝑛 = lim𝑛→∞ −4𝑛4 = 0.

//

Example 9[Computing limits of sequences using dominant term analysis]:

𝑛2 (2𝑛+1)(5−3𝑛)
Does the sequence 𝑎𝑛 = 1+2𝑛4
converge or diverge?

Solution: The dominant (highest degree) term in the numerator is 𝑛2 (2𝑛)(3𝑛). The dominant (highest degree)
term in the denominator is 2𝑛4 .

𝑛2 (2𝑛)(3𝑛)
lim𝑛→∞ 𝑎𝑛 = lim𝑛→∞ = 3. //
2𝑛4

Theorem 4[limit of a sequence via limit of a function]:

Let {an }∞
n=m be a sequence and suppose that f(x) is a real-valued function for which 𝑓(𝑛) = 𝑎𝑛 for all

integers 𝑛 ≥ 𝑚,

If lim𝑥→∞ 𝑓(𝑥) = 𝐿, then lim𝑛→∞ 𝑎𝑛 = 𝐿.

Remark 1: The converse of Theorem 3 is NOT true. See Example 9.

6
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Example 10: Let 𝑎𝑛 = 𝑠𝑖𝑛(𝑛𝜋)) and 𝑓(𝑥) = 𝑠𝑖𝑛(𝜋𝑥). Then lim𝑛→∞ 𝑎𝑛 = 0 but lim𝑥→∞ 𝑓(𝑥) does not exist
(DNE).

Remark 2: This theorem enables us to use L’Hopital’s rule for limits of certain sequences.

𝑛2
Example 11[determining limit of a sequence with L’Hopital’s rule]: Does the sequence 𝑎𝑛 = converge or
2𝑛

diverge?
1 1,9 25 36 49 64
Solution: The first 8 terms of the sequence are , , 1, , , , , …. It appears that it will keep going
2 8 32 64 128 256

down — looks like it will converge to zero. L’Hôpital’s rule proves it.

𝑥2
Let 𝑓(𝑥) = 2𝑥
, x≥ 1.

𝑥2 2𝑥 2
Then, lim𝑥→∞ f(x) = lim𝑥→∞ 2𝑥 = lim𝑥→∞ 2𝑥 𝑙𝑛2 = lim𝑥→∞ 2𝑥 𝑙𝑛2.𝑙𝑛2 = 0.

𝑛2
Thus lim𝑛→∞ 2𝑛 = 0. //

Example 12 [determining limit of a sequence with L’Hopital’s rule]: Find the limit of the sequence 𝑎𝑛 =
8𝑛
√5𝑛 .
1
𝑙𝑛5𝑥
Solution: Define (𝑥) = √5𝑥 , x≥ 1. Notice that 𝑎𝑛 = 𝑓 (𝑛). √5𝑥 = 𝑒 8𝑥(𝑙𝑛5𝑥) . But lim𝑥→∞
8𝑥 8𝑥
8𝑥
is
1
∞ 𝑙𝑛5𝑥 8𝑥
indeterminate form of the type ∞. By L’Hopital’s rule , lim𝑥→∞ 8𝑥
= lim𝑥→∞ 𝑥
8
=0. lim𝑥→∞ √5𝑥 = 𝑒 0 = 1.

8𝑛
We conclude that √5𝑛 → 1. //

Theorem 5[Squeezing rule for limit of a sequence]:

If 𝑎𝑛 ≤ 𝑐𝑛 ≤ 𝑏𝑛 ∀𝑛 and lim𝑛→∞ 𝑎𝑛 = 𝐿 = lim𝑛→∞ 𝑏𝑛 , then lim𝑛→∞ 𝑐𝑛 = 𝐿.

𝑛!
Example 13 [determining limit of a sequence with Squeezing rule: Let 𝑎𝑛 = 𝑛𝑛 . Show that 𝑎𝑛 → 0.

∞
Solution: Even if the limit has determinate form of the type ∞ we cannot apply L’Hopital’s Rule to solve this

problem, because it involves factorials. (𝑥! is not defined when 𝑥 is not a whole number. ).

𝑛! 1
Note that 0 ≤ 𝑎𝑛 = 𝑛𝑛 ≤ ∀𝑛.
𝑛

7
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

1 𝑛!
Since lim𝑛→∞ 0 = 0 = lim𝑛→∞ , so lim𝑛→∞ = 0 by squeezing rule. //
𝑛 𝑛𝑛

Theorem 6[Substitution rule for limit of a sequence]: If 𝑓 is a continuous function and

lim𝑛→∞ 𝑎𝑛 = 𝐿, then lim𝑛→∞ 𝑓( 𝑎𝑛 ) = 𝑓(𝐿) = 𝑓( lim 𝑎𝑛 )
𝑛→∞

𝑛!
Example 14[determining limit of a sequence with Substitution rule]: Let 𝑏𝑛 = cos ( 𝑛 ). Show that 𝑏𝑛 → 1.
𝑛

𝑛!
Solution: Let 𝑎𝑛 = . . From example--- 𝑎𝑛 → 0. Since 𝑓(𝑥) = 𝑐𝑜𝑠 𝑥 is a continuous function, so 𝑓(𝑎𝑛 ) →
𝑛𝑛

𝑓(0) = 1.

Caution: If f is not continuous at L, the conclusion of the above theorem can fail.

Example 15 [counter example when f is not continuous in Theorem 6]:

1 𝑖𝑓 𝑥 ≥ 0
Let 𝑓(𝑥) = {
−1 𝑖𝑓 𝑥 < 0

1
Let 𝑎𝑛 = − 𝑛. Then lim𝑛→∞ 𝑓(𝑎𝑛 ) = −1 ≠ f(lim𝑛→∞ 𝑎𝑛 ) = 𝑓(0) = 1. //

4.1.3 Bounded and Monotonic Sequences

In this subsection we study methods for testing convergence of a sequence without necessarily going for
finding their limits.

Definition 4 [bounded sequence]:

A sequence {an } is said to be

(a) bounded above if there exists a real number M such that 𝑎𝑛 ≤ 𝑀 ∀𝑛

(b) bounded below if there exists a real number m such that 𝑚 ≤ 𝑎𝑛 ∀𝑛
(c) bounded if it is bounded above and bounded below.

Remark: A sequence {an } is bounded if and only if there exists a positive number L such that
|𝑎𝑛 | ≤ 𝐿 ∀𝑛
Justify why!

1. Assume the definition. Then ∃ 𝑚, 𝑀 such that 𝑚 ≤ 𝑎𝑛 ≤ 𝑀 ∀𝑛. Define 𝐿: = max{|𝑚|, |𝑀|}.
8
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Then

−𝐿 ≤ 𝑚 ≤ 𝑎𝑛 ≤ 𝑀 ≤ 𝐿 ∀𝑛
This implies that
|𝑎𝑛 | ≤ 𝐿 ∀𝑛

2. Assume the remark. Then −𝐿 ≤ 𝑎𝑛 ≤ 𝐿 for some L. Put 𝑚: = −𝐿 and 𝑀: =L. Then {an } is bdd below
and above.

Definition 5 [monotone sequence]: A sequence {an }∞

n=m is said to be

a) increasing if 𝑎𝑛 ≤ 𝑎𝑛+1 ∀𝑛

b) decreasing if 𝑎𝑛 ≥ 𝑎𝑛+1 ∀𝑛

c) monotone if it is either increasing or decreasing.

d) Strictly increasing if 𝑎𝑛 < 𝑎𝑛+1 ∀𝑛

e) Strictly decreasing if 𝑎𝑛 > 𝑎𝑛+1 ∀𝑛

f) Strictly monotone if it is either strictly increasing or strictly decreasing

g) Eventually monotone if ∃a natural number N such that {an }∞

n=N is monotone.

9
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

sequences

increasing decreasing both

strictly strictly
increasing decreasing

non-strcictly non-strictly
decreasing decreasing

Remark 1 [Tests For Monotonicity of Sequences]: The following statements follow immediately from the
above definition.

(a) Difference test for monotonicity: The sequence{an }∞

n=m is increasing if and only if 𝑎𝑛 − 𝑎𝑛+1 ≤

0 and decreasing if and only if𝑎𝑛 − 𝑎𝑛+1 ≥ 0.

𝑎𝑛
(b) Ratio test for monotonicity: The sequence{an }∞
n=m is increasing if and only if 𝑎 ≤ 1 and decreasing if
𝑛+1
𝑎𝑛
and only if𝑎 ≥ 1.
𝑛+1

(1)(3)(5)…(2𝑛−1)
Example 16[determining monotone sequence by Ratio Test]: Let 𝑎𝑛 = (2)(4)(6)…(2𝑛)
, n=1,2,….

Show that the sequence {an }∞

n=1 is strictly decreasing.

𝑎𝑛+1 (1)(3)(5)…(2𝑛+1) (1)(3)(5)…(2𝑛−1) (2𝑛+1)

Solution: = (2)(4)(6)…(2𝑛+2) ÷ (2)(4)(6)…(2𝑛)
= (2𝑛+2) < 1. Hence, 𝑎𝑛+1 < 𝑎𝑛 ∀𝑛. Therefore, the sequence
𝑎𝑛

is strictly decreasing. //

Remark 2: We can use the derivative tests for functions to verify monotonicity of certain sequences.

3𝑛−1
Example 17 [determining monotone sequence by derivative Test]: Let 𝑎𝑛 : = 4𝑛+5, n=1,2,….
10
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Show that the sequence {an }∞

n=1 is strictly increasing.

3𝑥−1 19
Solution: Define 𝑓(𝑥) = , 𝑥 ∈ [1, ∞). Since 𝑓 ′ (𝑥) = (4𝑥+5)2 > 0 ∀𝑥, so 𝑓 is strictly increasing.
4𝑥+5

Therefore, 𝑓(𝑛) < 𝑓(𝑛 + 1). This implies 𝑎𝑛 < 𝑎𝑛+1 ∀𝑛. Hence, {an }∞
n=1 is strictly increasing. //

Theorem 7 [alteration of finite terms]: Let {𝑏1 , 𝑏2 , … , 𝑏𝑘 }be any finite sequence and N be any integer greater
than or equal to 𝑚.

∞
a) If {an }∞
n=m is convergent to L, then so are the sequences {a n }n=N and 𝑏1 , 𝑏2 , … , 𝑏𝑘 , 𝑎𝑚 , 𝑎𝑚+1 , ….

∞
b) If {an }∞
n=m is bounded, then so are the sequences {a n }n=N and 𝑏1 , 𝑏2 , … , 𝑏𝑘 , 𝑎𝑚 , 𝑎𝑚+1 , ….

Remark: Theorem 6(a) says adding or deleting a finite number of terms does not affect the limit of a sequence,
i.e., if a sequence is convergent, it remains convergent after a finite number of terms are added to it or deleted
from it. Similarly, Theorem 6/b/ says adding or deleting a finite number of terms does not affect the
boundedness of a sequence.

Theorem 8[a necessary condition for convergence]: If a sequence is convergent, then it is bounded.

Remark 1: the converse of Theorem 7 is NOT true, in general.

The theorem can be equivalently stated in the form of Remark 2 below.

Remark 2 [Test for Divergence of a sequence]: If a sequence is unbounded, then it is divergent.

Proof: This statement is a contrapositive of Theorem 8. ∎

Theorem 9[Test for convergence] or [A sufficient Condition for Convergence]:

a) Every increasing sequence which is bounded above is convergent (to the lub of the sequence).

b) Every decreasing sequence which is bounded below is convergent (to the glb of the sequence).

11
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Proof:

Remark 3: The above theorem is called the Monotone Convergence theorem for sequences of real numbers.

Example 19: Let 𝑎1 = 2, 𝑎𝑛 = √2 + 𝑎𝑛−1 for 𝑛 ≥ 2. Determine if this sequence is convergent and if it is
convergent, find its limit.

Solution: Observation of the first few terms of the sequence enables us to hypothesize the following
claim.

Claim:

(i) It is s increasing, i.e., 𝑎𝑛 ≤ 𝑎𝑛+1 ∀𝑛

(ii) It is bounded above by 2, i.e., , 𝑎𝑛 ≤ 2∀𝑛

Next we formally verify our guess (claim) by using the Principle of Mathematical Induction (P.M.I).

To prove (i) (by PMI):

Step 1: For n=1 the rule 𝑎𝑛 < 𝑎𝑛+1 is true, because 𝑎1 = 2 ≤ 𝑎2 = √2 + √2 .

Step 2: Assume the rule is true for 𝑛 = 𝑘, i.e., 𝑎𝑘 < 𝑎𝑘+1 .

⟹ 𝑎𝑘 + 2 ≤ 𝑎𝑘+1 + 2.

⟹ √𝑎𝑘 + 2 ≤ √𝑎𝑘+1 + 2.

⟹ 𝑎𝑘+1 ≤ 𝑎𝑘+2 .

12
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Hence, the rule 𝑎𝑛 < 𝑎𝑛+1 is true for 𝑛 = 𝑘 + 1, too.

By the PMI, the rule 𝑎𝑛 ≤ 𝑎𝑛+1 is true for every 𝑛.

To prove (ii) (by PMI):

Step 1: For n=1 the rule 𝑎𝑛 ≤ 2 is true, because 𝑎1 = 2 ≤ 2.

Step 2: Assume the rule is true for 𝑛 = 𝑘, i.e., 𝑎𝑘 ≤ 2.

⟹ 𝑎𝑘 + 2 ≤ 4.

⟹ √𝑎𝑘 + 2 ≤ √4.

⟹ 𝑎𝑘+1 ≤ 2.

Hence, the rule 𝑎𝑛 ≤ 2 is true for 𝑛 = 𝑘 + 1, too.

By the PMI, the rule 𝑎𝑛 ≤ 2 is true for every 𝑛.

By Theorem 8, the given sequence is convergent.

To find lim 𝑎𝑛 : Let lim𝑛→∞ 𝑎𝑛 = 𝑙. By Theorem 6, lim𝑛→∞ 𝑎𝑛−1 = 𝑙 .

By Theorem 5, lim𝑛→∞ 𝑎𝑛 = √2 + lim𝑛→∞ 𝑎𝑛−1

⟹ 𝑙 = √2 + 𝑙

⟹ 𝑙 = 2. //

4.1.4 Subsequences and Limit Points

Definition 6 [subsequence]: Let {an } be a sequence and {𝑛𝑘 }∞

𝑘=1 be a strictly increasing sequence of
∞
positive integers. Then the sequence {𝑎𝑛𝑘 } is called a subsequence of {an }.
𝑘=1

Example 20(subsequences and non-subsequences):

1 1
a. Let an = n, bn = 2n and cn = 1 for n=1,2,…Then {bn } is a subsequence of {an } but{cn } is not.

1 1 1 1 1 1 1 1 1 1
b. The sequence 4
, 2 , 6 , 8 , 10, . .. is not a subsequence of the sequence , 4 , 6 , 8 , 10,…
2

13
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Theorem 10[A Necessary Condition for convergence of a sequence]: If a sequence {𝑎𝑛 }∞

𝑛=1
∞
converges, then every subsequence {𝑏𝑘 }∞
𝑘=1 = {𝑎𝑛𝑘 } of the sequence converges to the same limit.
𝑘=1

Proof: Let 𝑎𝑛 → 𝑙 and let 𝑏𝑘 = 𝑎𝑛𝑘 where a strictly increasing sequence of positive integers. Let 𝜀 be an
arbitrary positive number. Since 𝑎𝑛 → 𝑙 , so here exists a natural number N such that |𝑎𝑛 − 𝐿| < 𝜀, ∀
𝑛 ≥ 𝑁. This implies

|𝑎𝑛 𝑘 − 𝐿| < 𝜀, ∀ 𝑛𝑘 ≥ 𝑁. Hence, |𝑏𝑘 − 𝐿| < 𝜀, ∀ 𝑘 ≥ 𝑁. [Note that 𝑘 ≥ 𝑁 implies 𝑛𝑘 ≥ 𝑁∀𝑘]

∎

Remark:

Corollary:

a. If {an } has two subsequence that converge to different limits, then ({an } does not converge.

b. If {an } has a divergent subsequence, then {an } does not converge

Definition7 [limit point, limit superior and limit inferior]:

Let {𝑎𝑛 }∞
𝑛=1 be a sequence and p a number.

We say p is a limit point of {an } if every open interval about 𝑝 contains an infinite number of an .

Note: There are two special limit points called limit superior and limit inferior are special limit points.

Proposition: Let {𝑎𝑛 }∞

𝑛=1 be a sequence. For each 𝑘, define 𝑠𝑘 ≔ sup{𝑎𝑘 , 𝑎𝑘+1 , … } and 𝑡𝑘 ≔

inf{𝑎𝑘 , 𝑎𝑘+1 , … }. Then, the sequences {𝑠𝑘 } and {𝑡𝑘 } are convergent.

Definition 8 [limit superior and limit inferior]:

Let {𝑎𝑛 }∞
𝑛=1 be a sequence. Then

a. the limit superior of {𝑎𝑛 }∞

𝑛=1 , denoted by limsup𝑛→∞ 𝑎𝑛 , is defined by

limsup𝑛→∞ 𝑎𝑛 := lim𝑛→∞ 𝑠𝑛 where. 𝑠𝑛 ≔ sup{𝑎𝑛 , 𝑎𝑛+1 , … }

14
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

a. the limit inferior of {𝑎𝑛 }∞

𝑛=1 , denoted by liminf𝑛→∞ 𝑎𝑛 , is defined by liminf𝑛→∞ 𝑎𝑛 := lim𝑛→∞ 𝑡𝑛

where 𝑡𝑛 ≔ inf{𝑎𝑛 , 𝑎𝑛+1 , … }

Thus, limsup𝑛→∞ 𝑎𝑛 := lim𝑛→∞ [ 𝑠𝑢𝑝{𝑎𝑘 : 𝑘 ≥ 𝑛}] and limsup𝑛→∞ 𝑎𝑛 := lim𝑛→∞ [ 𝑖𝑛𝑓{𝑎𝑘 : 𝑘 ≥ 𝑛}]

Read also limit superior and limit inferior of functions.

Example 21: Consider the sequence 1, 1,

⏟, 2 ⏟
1, 2, 3, ⏟ 1, 2, 3, 4, 5,… Every natural number is a
1, 2, 3, 4, ⏟

limit point of this sequence. Moreover its limit inferior is 1 and its limit superior is infinity.

Theorem 11[The Monotone Subsequence Theorem]: Every sequence of real numbers has a
monotonic subsequence.

Proof:

Theorem 12(Bolzano-Weierstrass): Let {an } be a bounded sequence. Then there exists a subsequence
of {an } call it {𝑎𝑛𝑘 } that is convergent.

Proof:

Theorem 13[a sufficient condition for convergence]: If {an } is a bounded sequence such that every
convergent subsequence {𝑎𝑛𝑘 } of {an } converges to L then {an } converges to L.

Proof:

4.1.5 Cauchy sequences [Read]

Exercise 1.1

1. Test the following sequences for convergence. For those which converge, find the limit.
3 3
a) 𝑎𝑛 = √𝑛 + 2 − √𝑛 + 1
b) 𝑎𝑛 = tanh(𝑛)
5
c) 𝑎𝑛 = 6−a with 𝑎1 = 4.
n

15
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Write the first four terms of the sequence and determine whether it is convergent or
divergent. If the sequence converges, find its limit.

a. {sinh n} b. n 2
 n  n2  n
c.  n  c  1
c 

2. Find lim a n if it exists where

n

n
 3
a. a n  1   b. an  3 n  2  3 n  1
 n
n
 n2 1 n n
c. a n    d. a n  sin
 n  n 1 2

n 2 (n!)
e. an  tanh n f. a n 
(n  2)!
3. Determine whether the given sequence is monotonic or not, and convergent or not.

 3n   n! 
a.  n 
b.  n c. { cos n }
2  3  n 

  3n 
d. ln  
  n  1 
e. 3
n n
4 n
  1  2 
f. 1   
 n  

 ln( n  2)   3n 
g.   h.  2 
 n2   (n  1) 

4. Determine the convergence or divergence of the sequence {an} if

3n
ln n  1  1  3  5  . . .  (2n  1)
a. a n  b. 1  2  c.
n 1  n  n2

16
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

a
 1
1  1  
 n  , where a and b are constants and b  0 .
n
d. a n   2  k e. a n  b
k 0  1
1  1  
 n

0 if r 1
5. Using lim r n  
 if r 1
n 

1  13 
n
1  (1) n
find a. lim b. lim c. lim n 4 n  5 n
n  2 n  e 2n n

6. Let {an} be sequence with a1  2 and an1  2  an , n  1.

a. Show that {an} is convergent

b. Find lim a n
n

1 1 1
7. Let i. a n   . . .  . Show that lim a n  1 .
1 2 2  3 n(n  1) n 

1 1 1
ii. bn   . . .  . Show that lim bn  1 .
n 
n2 1 n2  2 n2  n

8. Using the limit theorems of sequence it can be show that if a n  0 and bn  is bounded then lim a n bn  0 .
n 

Use this result to show the following sequences converge to 0.

 cos n   1  1  (1) n 1    2 n  (1) n 

i.   ii.  2 ln 1   iii.  
 n 
2n
n  n!   e 

10. Find two limit points and subsequences, which converge to each of these points for
  1   n 
a. (1) n  2   b. sin 
  n   4 

17
 Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

==============================================================================
4.2 Infinite series
4.2.1 Definition of Infinite Series
A real series (or an infinite series of real numbers) is an “infinite sum (a sum of infinite terms)” of the form
𝑎1 + 𝑎2 + ⋯

The associative property of addition guarantees that a finite sum of numbers of the type
𝑎1 + 𝑎2 + ⋯ + 𝑎𝑛

gives the same result no matter how we group them. For instance, 1+(2+3)=(1+2)+3. This is not the case for an
infinite sum. For example, it is suggestive that [1+(-1)]+[1+(-1)]+….=0 where as 1+[(-1)+1]+[(-1)+1]+…=1.
Hence, we need a careful definition for an infinite sum.

Definition 1 [series]: Given a sequence of real numbers {𝑎𝑘 }∞

𝑘=𝑚 , an expression (formal sum) of the form

𝑎𝑚 + 𝑎𝑚+1 + ⋯

is called an infinite series (or just a series). We denote this expression by ∑∞

𝑘=𝑚 𝑎𝑘 . The numbers

𝑎𝑚 , 𝑎𝑚+1 , …are called the terms of the series.

1 1 𝑛
Example 1: The expressions ∑∞ 𝑛 ∞ ∞
𝑘=0(−1) , ∑𝑘=0 𝑛 and ∑𝑘=0 (2) are infinite series.

4.2.2 Convergence and Divergence of Series

Definition 2 [partial sum and sum of a series]: Let ∑∞

𝑘=𝑚 𝑎𝑘 be a series.

a. The number 𝑆𝑛 := 𝑎
⏟𝑚 + 𝑎𝑚+1 + 𝑎𝑚+𝑛−1 is called the nth partial sum of the given series.
𝑡ℎ𝑒 𝑠𝑢𝑚 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑖𝑟𝑠𝑡 𝑛−𝑡𝑒𝑟𝑚𝑠

It is denoted by ∑𝑚+𝑛−1
𝑘=𝑚 𝑎𝑘 .
b. The sequence {𝑆𝑛 }∞
𝑘=𝑚 is called the sequence of partial sums.

c. The number S: =lim𝑛→∞ 𝑆𝑛 is called the sum of the series (if it exists). In this case we write S=∑∞
𝑘=𝑚 𝑎𝑘 .

D. We say the series ∑∞

𝑘=𝑚 𝑎𝑘 converges (is convergent) if lim𝑛→∞ 𝑆𝑛 exists (is a real number); otherwise,

we say it diverges ( is divergent).

Remark 1: In the theory of series, the expression ∑∞
𝑘=𝑚 𝑎𝑘 is used for two purposes. The first purpose is as an

abbreviation for the expression 𝑎𝑚 + 𝑎𝑚+1 + ⋯. The second purpose is to represent the sum of convergent
series, i.e., lim𝑛→∞ 𝑆𝑛 .

18
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Remark 2: There are several non-equivalent approaches to the definition of infinite sums. The definition given
here is the most standard. A branch of mathematics, called summability theory deals with the general theory.

Example 2: Use the definition to determine if each of the following series converges. If convergent, find the
sum.

1 𝑛 𝑛 1 4
a) ∑∞
𝑘=0 (2) b) ∑∞
𝑘=0 ( 𝑛(𝑛+1)
) c) ∑∞
𝑘=0(−1)
𝑛
d) ∑∞
𝑘=0 𝑛
e) ∑∞
𝑛=1 𝑛2 +4𝑛+3

4.2.3 Special Series

A. Telescoping Series

Note: The Series in (b) is called a telescoping series. In general, a telescoping series is a series in which most of
the terms cancel in each of the partial sums, leaving only some of the first terms and some of the last terms.

B. Harmonic series

The series in (d) is called harmonic series. A harmonic series is so named because it has to do with music (Read
Wikipedia article about harmonic series).

1 1 1 1 1 1 1 1
solution (d): Observe that 𝑆2𝑛 :=1 + 2 + [3 + 4] + [5 + 6 + 7 + 8] + ⋯ + 2𝑛

1 1 1 1 1 1 1 1
≥1+ +[ + ]+[ + + + ]+⋯+ 𝑛
2 4 4 8 8 8 8 2

1
≥ 1 + 𝑛( )
2
𝑛
Put 𝑏𝑛 ≔ 𝑎 + 2. Since {𝑏𝑛 } is divergent, so {𝑆2𝑛 } is divergent (by Corollary (b) of Theorem 10 of Section 1.1).

Since {𝑆2𝑛 } is a subsequence of {𝑆𝑛 }, so {𝑆𝑛 } is divergent (by Theorem 2 section 1.1). //

C. Geometric Series

Definition 3 [Geometric series]: A geometric series is a series of the form ∑∞ 𝑘

𝑘=𝑚 𝑐𝑟 , where 𝑐 and 𝑟 are fixed

real numbers and ≠ 0 .

19
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

𝑎𝑘+1
If 𝑎𝑘 ≔ 𝑐𝑟 𝑘 ,then = 𝑟 for every 𝑘. Hence, the ratio of consecutive terms does not depend on the running
𝑎𝑘

index. Hence, we call 𝑟 the common ratio of the geometric series ∑∞ 𝑛

𝑘=𝑚 𝑐𝑟 .

Theorem 2 [Geometric Series Test)]: The geometric series ∑∞

𝑘=𝑚 𝑐𝑟
𝑘

𝑐𝑟 𝑚
a) converges with sum 𝑆 = if |𝑟| < 1.
1−𝑟

b) diverges if |𝑟| ≥ 1.

Remark: In words, the sum S of a convergent geometric series ∑∞ 𝑘

𝑘=𝑚 𝑐𝑟 is given by

𝑓𝑖𝑟𝑠𝑡 𝑡𝑒𝑟𝑚
𝑆=
1 − 𝑐𝑜𝑚𝑚𝑜𝑛 𝑟𝑎𝑡𝑖𝑜

Proof: For the geometric series ∑∞ 𝑘

𝑘=𝑚 𝑐𝑟 the nth partial sum 𝑆𝑛 is turns out to be

1 − rn
m
Sn = cr ( )
1−r

1−rn
Thus, 𝑆 = lim𝑛→∞ cr m ( 1−r ). This limit exists if and only if |𝑟| < 1 .∎

Example 3. Determine if each of the following series converges or not. For those which are convergent, find
their sum.

a) ∑∞
𝑛=1(−1)
𝑛

22𝑛−1 5(−1)𝑛
b) ∑∞
𝑛=1 d) ∑∞
𝑘=1 4𝑛
3𝑛

c) ∑∞
𝑘=1 9
2−𝑛 𝑛+1
4
𝑎𝑛+1
Solution (a): Let 𝑎𝑛 ≔ (−1)𝑛 . Since = −1 is a constant (does not depend on the running index n), so
𝑎𝑛

the given series is geometric. Since, r ≥ 1, so it is divergent.

22𝑛−1 𝑎𝑛+1 4
Solution (b): Let 𝑎𝑛 ≔ . Since = 3 is a constant (does not depend on the running index n), so the
3𝑛 𝑎𝑛

given series is geometric. Since, r > 1, so it is divergent.

5(−1)4 𝑎𝑛+1 1
Solution (c): Let 𝑎𝑛 ≔ . Since = − 4 is a constant (does not depend on the running index n), so the
4𝑛 𝑎𝑛

given series is geometric. Since, |𝑟| < 1, so it is convergent. //

20
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

13 1
Example 4. If the sum of an infinite geometric series is and the common ratio is − , then find the first term.
15 4

1 13
Solution: Given: 𝑟 = − 4 and 𝑆 = .
15

Required: first term, 𝑎1 =?

1 𝑎
Formula: 𝑆 = 1−𝑟 ,

13 𝑎1
⟹ 15 = 1 .
1−(− )
4

13
⟹ 𝑎1 = 12. //

Example 5 [geometric series word problem]: A side of a square is 12 cm. The midpoints of its sides are
joined to form an inscribed square, and this process is continued. Find the sum of the perimeters of the squares
if this process is continued without end (round answer to two decimal places). Answer: 96 + 48√2

2.2.3 Properties of Convergent series

Theorem 1 [combinations of convergent series]: If ∑∞ ∞

𝑘=1 𝑎𝑘 and ∑𝑘=1 𝑏𝑘 are convergent and 𝑐 is any

constant, then the series ∑∞ ∞ ∞ ∞

𝑘=1(𝑎𝑘 + 𝑏𝑘 ), ∑𝑘=1(𝑎𝑘 − 𝑏𝑘 ) and ∑𝑘=1 𝑐𝑎𝑘 are convergent. More over if ∑𝑘=1 𝑎𝑘 =

𝐴 and ∑∞
𝑘=1 𝑏𝑘 = 𝐵, then

a. sum rule: ∑∞
𝑘=1(𝑎𝑘 + 𝑏𝑘 ) = 𝐴 + 𝐵

b. difference rule: ∑∞
𝑘=1(𝑎𝑘 + 𝑏𝑘 ) = 𝐴 − 𝐵

c. constant multiple rule: ∑∞

𝑘=1 𝑐𝑎𝑘 = 𝑐𝐴

Example 6: Prove: If the series ∑∞ ∞ ∞

𝑘=1 𝑎𝑘 converges and ∑𝑘=1 𝑏𝑘 diverges, then ∑𝑘=1(𝑎𝑘 + 𝑏𝑘 ) diverges.

Solution: Suppose ∑∞ ∞
[(𝑎𝑘 + 𝑏𝑘 ) − 𝑎𝑘 ] = ∑∞
𝑘=1(𝑎𝑘 + 𝑏𝑘 ) converges. Then, by Theorem 1(b), ∑𝑘=1 ⏟ 𝑘=1 𝑏𝑘

converges. This contradicts the assumption that ∑∞ ∞

𝑘=1 𝑏𝑘 diverges. Hence, our supposition that ∑𝑘=1(𝑎𝑘 + 𝑏𝑘 )

converges is wrong.

Remark 1 [term by term product]: If the series ∑∞ ∞ ∞

𝑘=1 𝑎𝑘 and ∑𝑘=1 𝑏𝑘 converge, then the series ∑𝑘=1 𝑎𝑘 𝑏𝑘

converges but its sum has nothing to do with the sums of ∑∞ ∞

𝑘=1 𝑎𝑘 and ∑𝑘=1 𝑏𝑘 .

Remark 2 [Cauchy product]: There is another way to "multiply" two series together to form a new convergent
series, called their Cauchy product.

21
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Definition 3 [Cauchy Product of two series]: Given two series ∑∞ ∞

𝑘=0 𝑎𝑘 and ∑𝑘=0 𝑏𝑘 , write

𝑐𝑛 := ∑𝑛𝑘=0 𝑎𝑘 𝑏𝑛−𝑘
Then 𝑐𝑛 is called the Cauchy product of ∑∞ ∞
𝑘=0 𝑎𝑘 and ∑𝑘=0 𝑏𝑘 .

The Cauchy product of two convergent series may happen to be divergent!

Theorem 2 [Cauchy product of two real series (Merten’s theorem)]: If ∑∞ ∞

𝑘=0 |𝑎𝑘 | and ∑𝑘=0 |𝑏𝑘 | both

converge to A and B respectively, then ∑∞

𝑘=0 |𝑐𝑘 | converges to C: = AB.

The definition of Cauchy product and Theorem 2 can be extended to series in ℂ and ℝ𝑛 .

Example 7[combining convergent series]: Determine the value of the following series.

∞ 4 ∞
∑ 2
− ∑ 92−𝑛 4𝑛+1
𝑛=1 𝑛 + 4𝑛 + 3 𝑛=1

Theorem 3 [alteration of finite terms of a series and re-indexing]:

1. Let 𝑚1 and 𝑚2 be any two whole numbers. Then the series ∑∞

𝑘=𝑚1 𝑎𝑘 converges if and only if the series

∑∞
𝑘=𝑚2 𝑎𝑘 converges. (alteration of terms)
2. It is always possible to regard ∑∞ ∞
𝑘=𝑚 𝑎𝑘 as ∑𝑛=0 𝑏𝑛 (i.e., as a series with initial index 0.). (re-indexing)

One useful consequence of Theorem 3a is that when investigating convergence of a series, we do not need to
care where the indexing starts. But when finding the sum we should care about where the index starts. When
investigating convergence of a series, we do not use full notation including index specification, we just write
∑ ak. A typical question then reads: "Is the series ∑ ak convergent?"

4.2.4 Convergence Tests for Positive/Non-negative Term Series

Convergence tests are theorems are methods of testing the convergence or divergence of series. A list of
common convergence tests are given below; but we learn only some of them in detail.

- Limit of the summand nth-term test, test for - p-series test

divergence, or the divergence test. - Direct comparison test
- Ratio test d'Alembert's criterion. - Limit comparison test
- Root test[edit]nth root test or Cauchy's - Cauchy condensation test
criterion. - Abel's test
- Integral test - Absolute convergence test
22
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

- Alternating series test - Weierstrass M-test

- Dirichlet's test - Extensions to the ratio test
- Cauchy's convergence test - Raabe–Duhamel's test
- Stolz–Cesàro theorem
- Bertrand's test - Gauss's test
- Kummer's test
Except divergence test all the test in this subsection are applicable only to either positive term or nonegative
term series.

A. Divergence Test; nth term test

Theorem 4[nth term Test]: If ∑∞

𝑘=1 𝑎𝑘 converges, then 𝑎𝑘 → 0.

1 1
Remark: The converse of Theorem is not true, in general. Counter example: 𝑘 → 0, but ∑∞
𝑘=1 𝑘 is

divergent.

Proof: Assume that ∑∞

𝑘=1 𝑎𝑘 = 𝑠. Let 𝑆𝑛 = 𝑎1 + ⋯ + 𝑎𝑛 and 𝑆𝑛−1 = 𝑎1 + ⋯ + 𝑎𝑛−1 .

Then lim𝑛→∞ 𝑎𝑛 = lim𝑛→∞ [𝑎1 + ⋯ + 𝑎𝑛 ] − [𝑎1 + ⋯ 𝑎𝑛−1 ] = lim𝑛→∞ [𝑆𝑛 − 𝑆𝑛−1 ] = 𝑠 − 𝑠 = 0. ∎

The next corollary is the contra positive of the theorem. It can be used to test the divergence of a series.
Corollary [Divergence Test]: If lim 𝑎𝑘 ≠ 0, then ∑∞
𝑘=1 𝑎𝑘 diverges.

Example 8 [on divergence test]:

𝑘 2 +3𝑘−1 𝑘 2 +3𝑘−1
a) The series ∑∞
𝑘=1 5𝑘 2 +7𝑘+10 is divergent, since 5𝑘 2 +7𝑘+10 does not converge to 0.

1 1
b) The series ∑∞
𝑛=1 𝑛𝑠𝑖𝑛(𝑛) diverges since 𝑛𝑠𝑖𝑛(𝑛) does not converge to 0.

B. Integral Test; p-test

Theorem 5 [Integral Test for series]: Let ∑∞

𝑘=1 𝑎𝑘 be a positive term series. Let there exists a function

f(x) such that the following conditions are satisfied:

i) 𝑓 is positive, continuous and decreasing on [1, ∞)

ii) 𝑎𝑛 = 𝑓(𝑛)∀𝑛

Then
23
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

∞
a) If ∑∞
𝑘=1 𝑎𝑘 converges, then ∫1 𝑓(𝑥)𝑑𝑥 converges.

∞
b) If ∫1 𝑓(𝑥)𝑑𝑥 converges, then ∑∞
𝑘=1 𝑎𝑘 converges

Proof (a): Let’s estimate the area under the curve 𝑦 = 𝑓(𝑥) on the interval [1, 𝑛] by taking rectangles of width
one and whose height is the right endpoint. This gives the following figure.

Figure 2: Illustrating Integrals Test for Convergence of Series

Now, note that,

𝑓(2) = 𝑎2 , 𝑓(3) = 𝑎3 , ⋯ , 𝑓(𝑛) = 𝑎𝑛

The approximate area is then,

𝐴 ≈ (1)𝑓(2) + (1)𝑓(3) + ⋯ + (1)𝑓(𝑛) = 𝑎2 + 𝑎3 + ⋯ + 𝑎𝑛

and we know that this underestimates the actual area so,

𝑛 n
∑ 𝑎𝑘 = 𝑎2 + 𝑎3 + ⋯ + 𝑎𝑛 < ∫ f(x)dx
𝑘=2 1

∞ ∞
Now, assume that ∫1 𝑓(𝑥)𝑑𝑥 is convergent and so ∫1 𝑓(𝑥)𝑑𝑥 must have a finite value. Also, because 𝑓(𝑥) is
positive we know that,

24
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

∞
n
∫ f(x)dx < ∫ 𝑓(𝑥)𝑑𝑥
1
1

This in turn means that,

∞
𝑛 n
∑ 𝑎𝑘 <∫ f(x)dx < ∫ 𝑓(𝑥)𝑑𝑥
𝑘=2 1
1

It follows that the nth partial sum 𝑆𝑛 = 𝑎1 + ∑𝑛𝑘=2 𝑎𝑘 of the series ∑∞

𝑘=1 𝑎𝑘 is bounded above by 𝑎1 +
∞
∫1 𝑓(𝑥)𝑑𝑥 =: 𝑀.

Next, because the terms are positive we also know that, the sequence {𝑆𝑛 }∞
𝑛=1 is also an increasing sequence. It

follows from the monotone convergence theorem for sequences that, the sequence {𝑆𝑛 }∞
𝑛=1 is convergent and

hence the series ∑∞

𝑘=1 𝑎𝑘 is convergent. So, part (a) of the test is proven.

Proof (b) [Exercise]: ∎

Example 9 [Using the integral test]: Using the integral test, determine if each of the following series is
convergent or not.

1 𝑘 1
a. ∑∞
𝑘=1 b. ∑∞
𝑘=1 𝑘 2 +1 c. ∑∞
𝑘=1 𝑘
𝑘𝑙𝑛𝑘

Corollary [P series Test]: Let 𝑝 be a constant. Then the series ∑∞

𝑘=1 𝑎𝑘 converges if and only if p>1. Such a

series is called a p series.

Example 10 [p series test]: For each of the following series, determine whether it converges or diverges.
1 1 1 1
a. ∑∞
𝑘=1 𝑘 2 b. ∑∞
𝑘=1 𝑘 4 c. ∑∞
𝑘=1 𝑘 2/3 d. ∑∞
𝑛=1 5
𝑛4

Note:

1
1. The exact value of ∑∞ th
𝑘=1 𝑘 2 baffled mathematicians until Leonhard Euler obtained it in 18 c.

1
2. Today the sum of the p series ∑∞
𝑘=1 𝑛𝑝 is known for every positive even integer p but not for a single odd

integer greater than 1.

25
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

1
3. In particular, even though the series ∑∞
𝑘=1 is known to converge, its sum is unknown (although it has
𝑘3

been proved by a French mathematician named Apery that the sum is irrational). But whether the sum
is algebraic or transcendental is yet to be proved.
4. Another important issue about the convergent p series is to find their approximate sums to any required
degree of accuracy.

Theorem 6[Remainder Estimate from the Integral Test]: Let ∑∞

𝑘=1 𝑎𝑘 be a convergent series with positive

terms. Let there exists a function f(x) such that the following conditions are satisfied:

i) 𝑓 is positive, continuous and decreasing on [1, ∞)

ii) 𝑎𝑛 = 𝑓(𝑛)∀𝑛

Let 𝑆𝑁 be the nth partial sum of ∑∞ ∞

𝑘=1 𝑎𝑘 and let 𝑅𝑁 : = ∑𝑘=1 𝑎𝑘 − 𝑆𝑁 . Then

∞ ∞

∫ 𝑓(𝑥)𝑑𝑥 < R N < ∫ 𝑓(𝑥)𝑑𝑥

𝑁+1 𝑁

1
Example 11: Determine the least value of 𝑁 necessary such that 𝑆𝑁 will estimate ∑∞
𝑘=1 𝑘 3 to within 0.001.

Answer: 23

C. Direct Comparison Test

Theorem 7 [comparison test for series]: Assume that 0 ≤ 𝑎𝑛 ≤ 𝑏𝑛 for all positive integer 𝑛.

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

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

Proof: Let 𝑆𝑛 := ∑𝑛𝑘=1 𝑏𝑘 and 𝑇𝑛 := ∑𝑛𝑘=1 𝑏𝑘 . Because 𝑎𝑛 ≥ 0&𝑏𝑛 ≥ 0, so the sequences {𝑆𝑛 }&{𝑇𝑛 } of both
partial sums form increasing sequences. Also, because 𝑎𝑛 ≤ 𝑏𝑛 for all n, we have 𝑆𝑛 ≤ 𝑇𝑛 ∀𝑛.

26
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Proof (a): Assume that ∑∞ ∞

𝑛=1 𝑏𝑛 converges. Thus there exists a real number 𝑇 such that ∑𝑛=1 𝑏𝑛 =

lim𝑛→∞ 𝑇𝑛 = 𝑇 . Since 𝑏𝑛 ≥ 0, so 𝑇𝑛 ≤ 𝑇. This implies that 𝑆𝑛 ≤ 𝑇. Hence, {𝑆𝑛 } is bounded above. Therefore,
from the monotone convergence theorem on sequences we know {𝑆𝑛 } is a convergent sequence and so
∑∞
𝑛=1 𝑎𝑛 converges. ∎

Proof (b): Exercise.

Example 11[Using the Comparison Test]: For each of the following series, use the comparison test to
determine whether the series converges or diverges.

1
(a) ∑∞
𝑛=1 𝑛3 +3𝑛+1 ans. convergent
1
(b) ∑∞
𝑛=1 ans. convergent
2𝑛 +1
1
(c) ∑∞
𝑛=1 𝑙𝑛𝑛 ans. Divergent
1
(d) ∑∞
𝑛=1 ans. Convergent
3𝑛 +𝑛
𝑛
(e) ∑∞
𝑛=1 ans. Divergent
𝑛2 −cos(𝑛2 )

D. Limit Comparison Test

𝑎𝑛
Theorem 8 [Limit comparison test for series]: Let 𝑎𝑛 , 𝑏𝑛 > 0 for all n≥1 and let lim𝑛→∞ 𝑏𝑛 = 𝐿.

a) If L is a positive number and ∑∞ ∞

𝑛=1 𝑏𝑛 converges, then ∑𝑛=1 𝑎𝑛 converges.

b) If L is a positive number and ∑∞ ∞

𝑛=1 𝑏𝑛 diverges, then ∑𝑛=1 𝑎𝑛 diverges.

c) If L=0 and ∑∞ ∞
𝑛=1 𝑏𝑛 converges, then ∑𝑛=1 𝑎𝑛 converges.

d) If 𝐿 = ∞ and ∑∞ ∞
𝑛=1 𝑏𝑛 diverges, then ∑𝑛=1 𝑎𝑛 diverges.

𝑎𝑛 𝑎
Proof (c): Suppose lim𝑛→∞ 𝑏𝑛 = 0. In this case, the sequence {𝑏𝑛} is bounded. As a result, there exists a
𝑛

constant M such that 𝑎𝑛 ≤ 𝑀𝑏𝑛 . Therefore, if ∑∞

𝑛=1 𝑏𝑛 converges, then ∑∞
𝑛=1 𝑎𝑛 converges.

Proof (d): On the other hand, suppose

27
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

𝑎𝑛 𝑎
lim𝑛→∞ = ∞. In this case, { 𝑛} is an unbounded sequence. Therefore, for every constant M there exists an
𝑏𝑛 𝑏𝑛

integer N such that 𝑎𝑛 > 𝑀𝑏𝑛 for all n≥N. Therefore, if ∑∞ ∞

𝑛=1 𝑏𝑛 diverges, then ∑𝑛=1 𝑎𝑛 diverges as well.

Remark: When applying the Limit Comparison Test to ∑∞

𝑛=1 𝑎𝑛 where 𝑎𝑛 is a rational function of n, you may

choose 𝑏𝑛 as follows:

dominating term in the numerator of an

bn =
dominating term in the dinominator of an

Example 12[Using the Limit Comparison Test]: Determine if the following series are convergent or not by
using the comparison test.

1
a. ∑∞
𝑛=1 𝑛2 −1 ans. convergent
1
b. ∑∞
𝑛=1 ans. Convergent
√𝑛+1
𝑙𝑛𝑛
c. ∑∞
𝑛=1 𝑛2

E. Ratio Test

𝑎𝑛+1
Theorem 9 [Ratio Test]: Let 𝑎𝑛 > 0 for all n≥1 and let lim𝑛→∞ 𝑎𝑛
= 𝑟.

a) if 0 ≤ 𝑟 < 1, then ∑∞
𝑛=1 𝑎𝑛 converges.

b) if 1 < 𝑟 ≤ ∞, then ∑∞
𝑛=1 𝑏𝑛 converges.

c) if 𝑟 = 1, the test is inconclusive.

Proof (a): Since 0≤ 𝑟 <1, there exists R such that 0 ≤ 𝑟 < 𝑅 < 1. Let = 𝑅 − 𝑟 . By the definition of limit of a
sequence, there exists some integer N such that

𝑎𝑛+1
| − 𝑟| < 𝜀 ∀𝑛 ≥ 𝑁. Thus,|𝑎𝑁+1 | < 𝑅𝑎𝑁 , |𝑎𝑁+2 | < 𝑅 2 𝑎𝑁 , …,|𝑎𝑁+𝑛 | < 𝑅 𝑛 𝑎𝑁 . Since R<1, the geometric
𝑎𝑛

series

28
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

𝑅|𝑎𝑁 |+𝑅 2 |𝑎𝑁 | + 𝑅 3 |𝑎𝑁 | + ⋯ converges. Using the comparison test, we can conclude that the series
|𝑎𝑁+1 |+|𝑎𝑁+2 | … converges. Therefore, since ∑∞ 𝑁 ∞ 𝑁
𝑛=1 |𝑎𝑛 |= ∑𝑛=1 |𝑎𝑛 | + ∑𝑛=𝑁+1 |𝑎𝑛 | where |∑𝑛=1 |𝑎𝑛 | is a

finite sum and ∑∞ ∞ ∞

𝑛=𝑁+1 |𝑎𝑛 | converges, we conclude that ∑𝑛=1 |𝑎𝑛 | (which means ∑𝑛=1 𝑎𝑛 ) converges. ∎

Proof (b): Exercise

Example 13[Using the Ratio Test]: Use ratio test to test the following series for convergence.

𝑛5
a) ∑∞
𝑛=1 ANS. convergent
5𝑛
2𝑛
b) ∑∞
𝑛=1 ANS. convergent
𝑛!
𝑛𝑛
c) ∑∞
𝑛=1 𝑛! ANS. divergent
(−1)𝑛 (𝑛)!
d) ∑∞
𝑛=1 (2𝑛)!
ANS. convergent

F. Root Test

Theorem 10 [Root Test]: Let 𝑎𝑛 ≥ 0for all n≥1 and let lim𝑛→∞ 𝑛√𝑎𝑛 = 𝑟.

a) if 0 ≤ 𝑟 < 1, then ∑∞
𝑛=1 𝑎𝑛 converges.

b) if 1 < 𝑟 ≤ ∞, then ∑∞
𝑛=1 𝑏𝑛 converges.

c) if 𝑟 = 1, the test is inconclusive.

Proof: Exercise.

The root test is useful for series whose terms involve exponentials. In particular, for a series whose
terms 𝑎𝑛 satisfy 𝑎𝑛 = 𝑏 𝑛 , we need only evaluate lim𝑛→∞ 𝑏𝑛 .

Example 14 [Using the Root Test]: Use ratio test to test the following series for for convergence.

𝑛
(𝑛2 +3𝑛)
a) ∑∞
𝑛=1 (4𝑛2 +5)𝑛 ANS. converge
𝑛𝑛
b) ∑∞
𝑛=1 (ln(𝑛))𝑛 ANS. divergent

29
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

4.2.5 Alternating series Test

Definition 3 [alternating series]: Any series of the form ∑∞

𝑘=1(−1)
𝑘+1
𝑏𝑘 or ∑∞ 𝑘
𝑘=1(−1) 𝑏𝑘 where 𝑏𝑘 ≥ 0 for

all positive integers k, is called an alternating series.

1 1 nCos(𝑛𝜋)
Example 15: The series ∑∞
𝑘=1(−1)
𝑘+1
𝑘
, ∑∞ 𝑘 ∞
𝑘=1(−1) 𝑘 and ∑𝑛=2 (𝑛−1)(𝑛+1) are alternating series. But the

sin(𝑛)
series ∑∞
𝑛=1 is neither nonnegative nor alternating.
𝑛(𝑛+1)

Theorem 11[Alternating Series Test]: An alternating series of the form ∑∞

𝑘=1(−1)
𝑘+1
𝑏𝑘 or
∑∞ 𝑘
𝑘=1(−1) 𝑏𝑘 converges if

a) lim𝑛→∞ 𝑏𝑛 = 0 and

b) the sequence {𝑏𝑛 }∞

𝑛=1 is decreasing

Remark: The converse of Theorem 4 is not true.

Example 16[Using Alternating Series Test]: Determine whether each of the following series is convergent or
divergent.

(−1)𝑛
a) ∑∞
𝑛=1 𝑛
Answer: convergent
(−1)𝑛−1
b) ∑∞
𝑛=1 2𝑛+5
ANS. convergent
𝑛
c) ∑∞
𝑛=1(−1)
𝑛−1
3𝑛−5
ANS. divergent
ln(𝑛)
d) ∑∞
𝑛=1(−1)
𝑛−1
ANS. convergent
𝑛

1 (−1)𝑛
Note: Recall that the harmonic series ∑∞ ∞
𝑛=1 𝑛 is divergent; but the alternating harmonic series ∑𝑛=1 is
𝑛
(−1)𝑘+1
convergent. In fact, we shall prove later in 2.3 that ∑∞
𝑘=1 𝑘2
= ln(2).

Theorem 12[Alternating Series Error estimate Test]: Assume that an alternating series of the form

∑∞
𝑘=1(−1)
𝑘+1
𝑏𝑘 or ∑∞ 𝑘
𝑘=1(−1) 𝑏𝑘 ,

30
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

satisfies the hypotheses of alternating series test. Let S denote the sum of the series, 𝑆𝑁 denote the Nth partial
sum of the series and let or 𝑅𝑁 = 𝑆 − 𝑆𝑁 . Then

|𝑅𝑁 | ≤ 𝑏𝑁+1

In other words, if the conditions of the alternating series test apply, then the error in approximating the infinite
series by the Nth partial sum 𝑆𝑁 is in magnitude at most the size of the next term 𝑏𝑁+1 .

Example 17 [Alternating Series Error Estimate]: Consider the alternating series

∞ (−1)𝑘+1
∑
𝑘=1 𝑘2

Use the remainder estimate to determine a bound on the error R10 if we approximate the sum of the series by the
partial sum S10.

Solution: From the theorem stated above,

1
|𝑅10 | ≤ 𝑏11 =
112

≈ 0.008265.

4.2.6 Absolute Convergence Test and Generalized Tests

We have not seen a test for the convergence of the series ∑∞

𝑘=𝑚 𝑎𝑘 where the terms are neither nonnegative nor

alternating, till now.

Theorem 13 [absolute convergence test]: If the series ∑∞ ∞

𝑘=1 |𝑎𝑘 | converges, then the series ∑𝑘=1 𝑎𝑘

converges, too.

Proof: Assume ∑∞ ∞
𝑘=1 |𝑎𝑘 | converges. By Theorem 1(c), ∑𝑘=1 2|𝑎𝑘 | also converges. Since 0 ≤ |𝑎𝑘 | + 𝑎𝑘 ≤

2|𝑎𝑘 |∀𝑘 and ∑∞ ∞

𝑘=1 2|𝑎𝑘 | is convergent, so by comparison test ∑𝑘=1 𝑎𝑘 + |𝑎𝑘 | converges, too. By Theorem

1(b), ∑∞ [𝑎𝑘 + |𝑎𝑘 | − |𝑎𝑘 | = ∑∞

𝑘=1 ⏟ 𝑘=1 𝑎𝑘 is convergent. ∎

31
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Remark: The converse of the theorem is NOT true. Counter Example: the alternating harmonic
(−1) 𝑘+1 1 𝑘+1 1 1
series∑∞ 𝑘 converges but the harmonic series ∑∞ |(−1) | = ∑∞ is divergent.
𝑘=1 𝑘=1 𝑘 𝑘=1 𝑘

Definition 3 [absolute and conditional convergence of a series]: We say ∑∞

𝑘=1 𝑎𝑘 converges

a) absolutely if ∑∞
𝑘=1 |𝑎𝑘 | converges

b) conditionally if∑∞ ∞
𝑘=1 𝑎𝑘 converges, but ∑𝑘=1 |𝑎𝑘 | diverges

Figure 3: Types of Infinite Series

Example 18 [Using absolute convergence test]: Determine convergence of the following series.

𝑠𝑖𝑛𝑛 (−1)𝑛
(a) ∑∞
𝑛=1 𝑛3 (c) ∑∞
𝑛=1 𝑛

(−2)𝑛 𝑛2 1−2𝑛 𝑛
(b) ∑∞
𝑛=1 (𝑛+1)!
(d) ∑∞
𝑛=1 (2+4𝑛)

Example 19 [absolute convergence, conditional convergence and divergence]: Test the following series for
absolute convergence, conditional convergence and divergence.

32
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

sin(𝑛)
(a) ∑∞
𝑛=1 𝑛(𝑛+1)
(−1)𝑛
(b) ∑∞
𝑛=2 𝑛(𝑛−1)
𝑛!(𝑛+1)!
(c) ∑∞
𝑛=1 (3𝑛)!
nCos(𝑛𝜋)
(d) ∑∞
𝑛=2 (𝑛−1)(𝑛+1)

33
 Theorem 14 [properties of absolutely convergent series]: If ∑∞ ∞
𝑘=1 𝑎𝑘 and ∑𝑘=1 𝑏𝑘 are

absolutely convergent and c is a constant, then ∑∞ ∞ ∞

𝑘=1(𝑎𝑘 + 𝑏𝑘 ), ∑𝑘=1(𝑎𝑘 − 𝑏𝑘 ) and ∑𝑘=1 𝑐𝑎𝑘

are absolutely convergent.

Example 20 [absolute convergence]

Theorem 15 [The Rearrangement Theorem for Absolutely convergent series]: If ∑∞

𝑘=1 𝑎𝑘

converges absolutely and {𝑏𝑘 } is any rearrangement of the terms the sequence {𝑎𝑘 } , then
∑∞ ∞
𝑘=1 𝑏𝑘 converges absolutely and both series and ∑𝑘=1 𝑏𝑘 converge to the same value.

The condition of absolute convergence of the series thus guarantees that any rearrangement of
terms is also convergent and the sum is unchanged. If the series is not absolutely convergent but
still conditionally convergent we have another remarkable result stated in a theorem formulated
by Riemann as follows.

Theorem 16 [The Rearrangement Theorem for conditionally convergent series]: If ∑∞

𝑘=1 𝑎𝑘

converges conditionally and L is a real number, ∞ or −∞, then there is a rearrangement {𝑏𝑘 } of
{𝑎𝑘 } such that ∑∞
𝑘=1 𝑏𝑘 = 𝐿.

4.2.7 Advanced Convergence Tests and a concluding remark

The list of convergence tests considered so far includes only commonly used tests for
convergence that can be found in most calculus books. More advanced tests of convergence
(such as Gauss test, raabe test, Kummer Test, etc) are given in advanced Calculus or Real
Analysis textbooks.
We now have several ways of testing a series for convergence or divergence; the problem
is to decide which test to use on which series. In this respect testing series is similar to
integrating functions. There are no hard and fast rules about which test to apply to a given series
but you may find the following flow charts helpful.
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Figure 4: Flow Chart No. 1of convergence-divergence tests

35
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Figure 5: Flow Chart No.2 of Convergence-divergence tests

36
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Exercise 2.2

1. What is the sum of an infinite geometric series if the first term is 2, and the common
ratio is -1/2?
2. The sum of an infinite geometric series is four times the first term. What is the
common ratio?
3. The common ratio in a geometric series is one third the first term. The sum of the
infinite series is 0.6. What is the first term?
4. The sum of an infinite geometric series is 24, and the sum of the first 200 terms of the
series is also 24. What are the first term and common ratio of the series?
5. A ball is thrown 12 meters in the air (so that the initial up-and-down distance is 24
meters). The ball rebounds 95% of the distance it falls. What is the total vertical
distance traveled by the ball before it stops bouncing? 480 meters
1 1 1 1
6. Find the limit as n→∞ of 𝑛 + 𝑛+1 + 𝑛+2 + ⋯ + 3𝑛
kp
7. For what values of p does the series ∑∞
k=1 2+k3 converge?

8. Find the sum S of the following series

12 1+2𝑘 4
a. ∑∞
𝑘=1 𝑘 2 +4𝑘+3 b. ∑∞
𝑘=1 3𝑘−1 c. ∑∞
𝑘=1 [𝑘 2 +4𝑘+3 − 9
−𝑘+2 𝑘+1
4 ].

9. Let S = {x1 , x2 , . . . } be the collection of all natural numbers that do not contain the
digit 0, i.e. 6∈S but 60∉S. Prove that ∑∞
k=1 xk <90. [Hint: Partition S into an infinite

collection of sets {S1,S2,...,Sj,...} where Sk⊂S is the set of all natural numbers that do
not contain the digit and have j digits total and use geometric series test]

37
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

4.3 Power Series

4.3.1 Definition of Power Series
Definition 1: Let x be a variable. A power series in x is a series of the form


c
n 0
n x n  c0  c1 x  c 2 x 2    c n x n  

where each ci is a real number.

In simplifying the nth term of a power series we always assume that x0 = 1 for every real
x including x = 0. The main objective of this section is to determine all values of x for which a
power series converges. Note that every power series in x converges if
x = 0, since
c0  c1 0  c 2 (0) 2    c n (0) n    c0 .

To find other values of x that produce convergent series, we often employ the Generalized Ratio
Test for absolute convergence.
Example 1. Show that


 n! x
n 0
n
 1  x  2x 2  6x3 

converges only for x = 0.

Solution: If x  0

(n  1)! x n1
lim  lim (n  1) x  
n n! x n n

Thus the Generalized Ratio Test implies that the series diverges.

Example 2. Find all the values of x for which the following power series is absolutely
convergent:
  n
xn  x
a.  2 3
b. 1 + x +2x + 3x + . . . c.  n 
n 0 n! n 0  3 

Solution: a. Using the Generalized Ratio Test, since

38
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

x n1 /( n  1)! 1
lim  lim x 0
n n
x / n! n n  1

we conclude that the series converges for every real number.

b. Since

(n  1) x n1
lim  x
n nx n
by the Generalized Ratio Test with r =|x|, the series is absolutely convergent if |x|<1
that is the power series converges on the interval -1<x<1 and diverges for x<-1 or
x>1.
If x=1 the series is 1+1+2+3+ . . . and is divergent.
If x=-1 the series is 1-1+2-3+ . . . and is divergent.
c. Since
n  1 n 1
x
n 1 n 1 1
lim 3  lim x  x
n  n n n  3n 3
n
x
3
1
by the Generalized Ratio Test with r = x , the series is absolutely convergent if
3
1
x <1 that is the power series converges on the interval –3<x<3 and the series
3
1
diverges if x >1 i.e x<-3 and x>3.
3
1
If x =1 the ratio test is inconclusive, and hence the number 3 and –3 require
3
special consideration, substituting 3 for x in the power series we obtain
0+1+2+3+ . . .+ n + . . .
which is divergent, by the nth term test because lim a n  0 . If we let x = -3
n 

we obtain
0-1+2- 3 + . . .+ (-1)nn + . . .
but since lim a n  0 by the nth term test the series is divergent.
n 

39
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

4.3.2 Convergence and divergence of Power Series ; Radius and Interval of

Convergence
The following theorem tells us that we can divide power series in to three categories.

Theorem 1: Let c
n 1
n x n be a power series, then exactly one or the following conditions holds.


a. c
n 1
n x n converges only for x=0.


b. c
n 1
n x n converges for all x.


c. There is a number R>0 such that c
n 0
n x n converges for |x|<R and diverges

for |x|>R.

We call the number R in part (c) of the above theorem the radius or convergence of the series.
If the series satisfies a) then we let R=0. If the series satisfies b) then we let R=  .
The collection of values of x for which the series converges is called the interval of
convergence of the series. From conditions a, b, and c of theorem 3.2 it is apparent that the
interval of convergence takes one and only one of the following forms:
[0,0], (-R,R), [-R,R), (-R,R], [-R,R], (, )
ILLUSTRATION
_____________________________________________________________________
Example Interval Radius of convergence
_____________________________________________________________________


 n! x
n 0
n
[0,0] 0

1  x  2 x 2  3x 3  . . . (-1,1) 1

xn

n 1 n
[-1,1) 1


 x n

n 1 n
(-1,1] 1

40
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)


 1n
 2n  1 x
n 0
2 n 1
[-1,1] 1

 n
 x
 n 
n 1  3 
(-3,3) 3


xn

n 0 n!
(, ) 

Class Work: Determine the interval of convergence of the power series

 
1 n2 n
1. n 1 n
xn 2. 
n 0 2
n
x

We shall also consider the following more general types of power series.

Definition 2.Let 𝑎 be a real number and 𝑥 a variable. A power series in x-a is a series of
theform


c
n 0
n ( x  a) n  c0  c1 ( x  a )  c 2 ( x  a) 2    c n ( x  a) n  

where each ak is a real number.

To simplify the nth term in (3.3), we assume that (x-a)0 =1 even if x = a. As in (3.2) x
replaced by x-a, exactly one of the following cases is true:
(i) The series converges only if x-a =0 that is, if x = a.
(ii) The series is absolutely convergent for every x.
(iii) There is a number R>0 such that the series is absolutely convergent if x is in the
open interval (a - R, a + R) and divergent if x < a - R or x >a + R.
As before, the totality of numbers for which the series converges is called the interval of
convergence, and R is the radius of convergence.

Example 3 Find the interval of convergence of the series


x  3n

n 1 n  4n
Solution: Since

41
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

R= lim
x  3n1  n4 n  lim
n x3 1
  x3
n  n  14 n 1 ( x  3) n n  n  1 4 4

1
by the Generalized Ratio Test the series is convergence absolutly for R= x  3 <1, that is, for –
4
4<x+3<4 or –7<x<1, and diverges for x<-7 and x>1.

 1n
For x = -7, the series is 
n 1 n
which is the negative of the convergent alternating series so


 1n
then the series 
n 1 n
converges.


1
For x =1, the series is the divergent harmonic series  n . Thus the interval of convergence is
n 1

given by [-7,1).
Class Work

( x  3) n
Find the interval of convergence of the series  (1) n
n 0 n 1
.

4.3.3 Differentiation and Integration



Since n 0
cn x n is a function, we ask whether its derivative exists. The next theorem gives us a

criterion for finding its derivative.

Theorem 2 (Differentiation Theorem for Power Series)

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

 
n 1
ncn x n1 has

the same radius of convergence, and

d  
 
 
d
 c n x    n 1 ncn x  
 n 1
 n
cn x n for |x|<R
dx  n 0  n 1 dx

Example 1 Show that

d   xn   xn
   
dx  n0 n!  n0 n!

42
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)


xn
Solution: We know from one of our examples that the series 
n 0 n!
converges for all x. The



Differentiation Theorem tells us that n 1
nx n1 / n! converges as well and that

d   x n   d  x n   nx n1 
x n1 
xn
           .
dx  n0 n!  n1 dx  n!  n1 n! n 1 (n  1)! n 0 n!

If we define a function f by

xn
f(x)=  for all x,
n 0 n!

then by example one above

f’(x)=f(x) for every x.

0n
Since f(0) =   1 , we conclude from the Exponential Growth Theorem i.e
n 0 n!

[Suppose f is continuous on [0, ) and f ' (t )  kf (t ) for t>0, then f (t )  f (0)e kt for t  0. ] that
f(x) = ex every x.
Therefore we may give the following definition

xn x2 x3
e= 
x
=1  x   
n 0 n! 2! 3!
Thus from the power series expansion of ex we have the following results.
1
 e 1  

 1 n 
e x  
 x n 

 1n x n
e n 0 n! n 0 n! n 0 n!

ex  
2

x 
2 n


x 2n
n 0 n! n 0 n!
Class Work
1. Using the power series expression for ex show that
 
1 2n 1
a. cosh x   x x   b. sinh x   x 2 n 1 x  
n  0 2n ! n  0 2n  1!

2. Use a power series representation for 1/(x+1) to obtain a power series representation r
1
if |x| < 1.
1  x 2

43
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

x
3. Find the power series representation for .
(1  x ) 2
Integration of power series

The next theorem states that we can also integrate the power series c
n 0
n x n term by term, just as

we did for differentiation.

Theorem 3 (Integration Theorem for Power Series)
 
cn
Let  cn x n be a power series with radius of convergence R>0, then
n 0
 (n  1) x
n 0
n 1
has the

sameradius of convergence, and

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

Example 2 Show that ln( 1  x)  


 1n x n 1  

 1n1 x n for |x| < 1
n 0 (n  1) n 1 n
Solution: We have seen that in class work (2) above that

1
  (1) n t n for |t|<1.
1  t n 0
Thus using integration theorem for power series we have
x x 
1
ln( 1  t )   dt   ( (1) n t n )dt
0
1 t 0 n 0

(1) n n 1
 x for x  1.
n 0 n  1

Therefore
x2 x3 x4
ln( 1  x)  x    
2 3 4
Example 3 Show that

(1) 2 n 1
arctan x   x for x  1 .
n  0 2n  1

Then the integration theorem yields

44
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

x  x
1
arctan x  
1  t 2
dt   (  (1) n t 2 n )dt
0 0 n 0

(1) n 2 n 1
 x for x  1.
n  0 2n  1

Therefore
x3 x5 x7
arctan x  x    
3 5 7

0.1

e
 x2
Example 4 Approximate dx.
0

Solution: We cannot use the fundamental theorem of calculus to evaluate the integral, because

we do not know the antiderivative of e  x . Although we could use other methods, the following
2

method is simpler and, in addition, produces a high degree of accuracy using the sum of only two
terms. Letting x  t 2 in the power series of ex, we get
2 t4 (1) n t 2 n
et  1  t 2   
2! n!
for every t. Applying the integration theorem, we find that
0.1 0.1
 e  x dx   e t dt
2 2

0 0
0.1 0.1
t 3  t5 
 t        
0.1
0
 3 0 10  0

(0.1) 3 (0.1) 5
=0.1-   .
3 10
If we use the first two terms to approximate the sum of this convergent alternating series, then,
by Theorem (2.27), the error is less than the third term (0.1)5/10=0.000001. Hence
0.1
0.001
e
 x2
dx  0.1   0.09967,
0
3
which is accurate to five decimal places.

45
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Class Work
1. Prove that

xn
ln( 1  x)   if x  1.
n 1 n
1
2

2. Use an infinite series to approximate the to four decimal places  arctan x 2 dx .

0

2.3.4 Taylor Series and its Application

The main advantage of having a power series representation of a function f on an open
interval I containing 0, is having f as the sum of a convergent series and hence we can
approximate it by partial sums. Since the partial sums of a power series are polynomials, values
of functions having power series representations are easily approximated.

Theorem 4: Suppose a power series c n 0
n x n has radius of convergence R>0. Let


f(x)= c
n 0
n x n for –R<x<R,

then f has derivatives of all orders on (-R,R), and

f n (0)  n!c n for n  0 .

Consequently,

f ( n ) (0) n
f ( x)   x for –R<x<R
n 0 n!

Definition 3: Suppose that f has derivatives of all orders at 0. Then the Taylor series of f is
the power series

f n (0) n

n 0 n!
x

46
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Definition 4: Let f be a function that has n derivatives at 0: The nth degree Taylor polynomial
Pn(x) of f at 0 is
f " (0) 2 f n (0) n
Pn ( x)  f (0)  f ' (0) x  x  x .
2! n!
For instance the 3rd degree Taylor polynomial P3(x) at 0 is given by
f " (0) 2 f (3) (0) 3
P3 ( x)  f (0)  f ' (0) x  x  x
2 6
If f is a function with higher derivatives at o, then in order to measure how well Pn(x)
approximates f(x), we define the nth Taylor remainder rn of f by the formula
rn ( x)  f ( x)  Pn ( x).
The value of rn(x) tells us how well Pn(x) approximates f(x).
It follows that the Taylor polynomials of f are the partial sums of the Taylor series of f and that

f n (0) n
f ( x)   x if and only if lim rn ( x)  0
n! n 
n 0

Theorem 5 (Taylor’s Theorem)

Let n be a nonnegative integer, and suppose f n1 ( x) exists for each x in an open interval I
containing 0. For each x  0 in I, there is a number tx strictly between 0 and x such that

f ( n) (0) n f ( n1) (t x ) n1

f ( x)  f (0)  f ' (0) x    x  x (1)
n! (n  1)!

f ( n1) (t x ) n1
and rn ( x)  x (2)
(n  1)!
(Equation (1) is known as Taylor’s Formula, and (2) is known as the Lagrange
RemainderFormula.)

Example 1 Show that

x2 x4 x6 
(1) n 2 n
cos x  1       x for every x.
2! 4! 6! n 0 (2n)!

47
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Solution: Let f ( x)  cos x , then

f ' ( x)   sin x f " ( x)   cos x f (3) ( x)  sin x f ( 4) ( x)  cos x

f (5) ( x)   sin x f ( 6) ( x)   cos x and so on.

Consequently
f ( 2 k 1) ( x)  (1) k 1 sin x hence f ( 2 k 1) (0)  0, and

f ( 2 k ) ( x)  (1) k cos x hence f 2 k (0)  (1) k .

Thus we suppress the odd powers and write the Taylor series of cosx as

x2 x4 x6 
(1) n 2 n
1      x
2! 4! 6! n 0 (2n)!

To show that the Taylor series converges to cos x for every x, we will show that lim rn ( x)  0 for
n 

all x. We can observe from our arguments above that, irrespective of the

( n 1)
integer n or the numbers x and tx, we have f (t x )  1, so that

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

it follows from one of our examples that

n 1 n
x x
lim  lim 0
n (n  1)! n n!
Therefore lim rn ( x)  0 for all x. Hence the Taylor series of the cosine function converges to
n 

cos x for every x.

The classic example below brings out the fact that a function f may have derivatives of all orders
at some number c, but may not have a Taylor series representation at that number. This shows
that an additional condition such as lim rn ( x)  0 , is required to guarantee the existence of a
n 

Taylor series.
Example 2 Let f be the function defined by

48
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)


1
e x2
if x  0
f ( x)  

0 if x  0

Show that f(x) cannot be represented by a Taylor series.

Solution:
1
f ( x)  f (0)
2
e x (1 / x)
f ' (0)  lim  lim  lim 1 / x 2 . Applying L’H ô pital’s rule, we see that
x 0 x0 x 0 x x 0
e
(1 / x 2 ) x
f ' (0)  lim 1/ x2
 lim 2
 0.
x 0
(2 / x )e 3 x 0
2e 1 / x

It can be proved that f " (0)  0, f " ' (0)  0, and, in general, f ( n ) (0)  0 for every positive
integer n. Then if f has a Taylor series expansion then
0 2 0
f ( x)  0  0 x  x    x 2  ,
2! n!
which implies that f ( x)  0 throughout an interval containing 0. However, this contradicts the
definition of f. Consequently f(x) does not have a Taylor series representation.
Class Work
Find the Taylor series representation of
a. sin x b. ln(1+x)

Taylor Series About an Arbitrary Point

We are interested in the possibility of expressing a function as a power series


c
n 0
n ( x  a) n (1)

in power or (x-a), where a can be fixed number. All the results we obtained for ordinary power
series have their counter part for power series of the type given in (1).In particular, if f has
derivatives of all orders at a, then we call

f n (a)

n 0 n!
( x  a) n (2)

the Taylor series of f about the number a. The nth Taylor polynomial P n of f about a is defined
by

49
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

f " (a) f (n)

Pn ( x)  f (a)  f ' (a)( x  a)  ( x  a) 2    ( x  a) n
2! n!
and the nth Taylor remainder rn of f about a is defined by
rn ( x)  f ( x)  Pn ( x)

Theorem 6: Let n be a nonnegative integer, and suppose f ( n1) ( x) exists for each x in an
openinterval I containing a. For each x in I, there is a number tx strictly between a
and x such that
f " (a) f ( n) f ( n1) (t x )
f ( x)  f (a)  f ' (a)( x  a)  ( x  a) 2    ( x  a) n  ( x  a) n1
2! n! (n  1)!

f ( n1) (t x )
and rn ( x)  ( x  a) n1
(n  1)!
(Where the first equation is known as Taylor’s series and the second equation is called
Lagrange Remainder Formula.)
Example 3 Express the polynomial f ( x)  2 x 3  9 x 2  11x  1 as polynomial in (x-2).
Solution: We wish to write f(x) in the form of (2) with a=2. To do so we must compute the
derivatives of f at 2:
f ( x)  2 x 3  9 x 2  11x  1 f (2)  1
f ' ( x)  6 x 2  18 x  11 f ' ( 2)   1
f " ( x)  12 x  18 f " (2)  6
f ( 3)
( x)  12 f ( 3) (2)  12
f ( n ) ( x)  0 f ( n ) ( 2)  0 for n  4
Therefore
6 12
f ( x)  1  (1)( x  2)  ( x  2) 2  ( x  2) 3
2! 3!
 1  ( x  2)  3( x  2)  2( x  2) 3
2


Example 4 Find the Taylor series of sin x about , and show that it converges to sin x for all
6
x.

f n (a) 
Solution: Let f ( x)  sin x . We wish to write f(x) as n 0 n!
( x  a) n where a  . Then the
6
derivatives of f are

50
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

f ( x)  sin x f ( / 6)  1 / 2

f ' ( x)  cos x f ' ( / 6)  3 / 2

f " ( x)   sin x f " ( / 6)  1 / 2
f 3 ( x)   cos x f 3 ( / 6)   3 / 2
f 4 ( x)  sin x f 4 ( / 6)  1 / 2

Since f 4 ( x)  f ( x), and as the absolute value of the inbetween derivative values are less than
one for every real number x we deduce that
( n 1)
f (t x )  1 for all real number x.

and thus
f ( n1) (t x )  n1 ( x   / 6) n1
rn ( x)  (x  ) 
(n  1)! 6 (n  1)!

since lim
x   / 6n1  n!
 lim
1
x  /6  0
n  (n  1)! x   / 6 n n  n  1

we conclude from corollary (2.18) that

( x   / 6) n1
lim 0
n (n  1)!
Then it follows that lim rn ( x)  0 for all real number x by squeezing theorem.
n 

As the result, the Taylor series of f converges to f. Moreover

f ( 4 n ) ( / 6)  1 / 2, f ( 4 n 1) ( / 6)  3 / 2
f ( 4 n  2 ) ( / 6)  1 / 2 f ( 4 n 3) ( / 6)   3 / 2
Therefore

 1 1  
2 3
1 3 3 1
f ( x)   x  x     x   
2 2  6  2 2!  6 2 3!  6

  1  
2 3
1 3 3
f ( x)   x   x     x   
2 2  6 4 6 12  6
The above equation in condensed form is given by

1   1 3 
f ( x)  
2 n 0
(1) n  ( x   / 6) 2 n 
2(n  1)!
( x   / 6) 2 n 1 .
 (2n)! 
Class work

51
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Find the Taylor series of f about a. In each case show that the Taylor series you found converges
to f by demonstrating that lim rn ( x)  0 for all x.
n 

a. 4x2-2x+1; a=0,-3. b. 5x3 +4x2+3x+2;a=0,2 c. cos x; a   / 3.

Application of Taylor series
Given a function f that satisfies the hypotheses of Taylor’s formula with remainder, we
have
f ( x)  Pn ( x)  rn ( x),
where Pn(x) is the nth-degree Taylor polynomial of f at a and rn(x) is the Taylor
remainder. If lim rn ( x)  0 , then, as n increases, Pn(x)  f (x ) ; hence the approximation
n 

formula f ( x)  Pn ( x) improves as n gets larger. Thus, we can approximate values of many

different transcendental functions by using polynomial functions. This is a very important fact,
because Polynomial functions are the simplest functions to use for numerical evaluation of
functions.
For instance consider the exponential function given by f ( x)  e x . The Taylor series
representation of ex is
x2 x3 xn
ex  1 + x +   
2! 3! n!
If we approximate e x by means of Taylor polynomials (with a=0), we obtain
e x  P1 ( x)  1 + x

x2
e x  P2 ( x)  1  x 
2
x2 x3
e x  P3 ( x)  1  x  
2 6
and so on. These approximation formulas are accurate only if x is close to 0. To approximate e x
for larger values of x, we may use Taylor polynomials with a  0.
In the next two examples, Taylor polynomials are used to approximate function values. If
we are interested in k-decimal-place accuracy in the approximation of a sum, we often
approximate each term of the sum to k+1 decimal places and then round off the final result to k
decimal places.

52
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Example 5 Let f ( x)  ln x.
a. Find P3(x) and r3 ( x) at c=1.

b. Approximate ln 1.1 to four decimal places by means of P3(1.1), and use r3 (1.1) to
estimate the error in this approximation.
Solution:
a. The general Taylor polynomial P3(x) about c=1 and Taylor remainder r3 ( x) at c=1 are

f ' (1) f " ( x)

P3 ( x)  f (1)  f ' (1)( x  1)  ( x  1) 2  ( x  1) 3 and
2! 3!
f ( 4) (t x )
r3 ( x)  ( x  1) 4
4!
Where tx is a number between 1 and x. Thus, we need the first four derivatives of f. It is
convenient to arrange our work as follows;
f ( x)  ln x f (1)  0
1
f ' ( x)  f ' (1)  1
x
f " ( x)   x  2 f " (1)  1
3
f ( 3)
( x)  2 x f ( 3) (1)  2
4
f ( 4 ) ( x)  6 x  4 f ( 4 ) (t x )  6t x

Substituting in P3(x) and r3 ( x) , we obtain

1 2
P3 ( x)  0  ( x  1)  ( x  1) 2  ( x  1) 3
2! 3!
4
 6t x 1
r3 ( x)  ( x  1) 4   4
( x  1) 4
4! 4t x
Where tx is a number between 1 and x.
a. From part (a)
1 1
ln 1.1  P3 (1.1)  0.1  (0.1) 2  (0.1) 3  0.0953
2 3
To estimate the error in this approximation, we consider

(0.1) 4 0.0001
r3 (1.1)   4
   0.000025.
4t x 4

Therefore ln 1.1  0.0953 accurate to four decimal places.

53
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Example 6 Use a Taylor polynomial to approximate cos 61 ,and estimate the accuracy of the
approximation.
 
Solution: We wish to approximate f ( x)  cos x if x  61. Put x  60   1   and
3 180

choose a  the choice of n will depend on the degree of accuracy we wish to attain. Let us try
3
n  2. Then
  1
f ( x)  cos x f 
3 2
  3
f ' ( x)   sin x f '   
3 2
  1
f " ( x)   cos x f "   
3 2
f (3) ( x)  sin x f (3) t x   sin t x

The second-degree Taylor polynomial of at a 
3

  1 
2
1 3
P2 ( x)   x  x   .
2 2  3 4 3

Since x represents a real number, we convert 61 to radian measure before substituting into
P2 ( x) . Writing

  3     1   
2
1
P2 (  )          or
3 180 2 2  3 180 3  4  3 180 3 
 
3   1  
2
1
P2 (  )       0.48481
3 180 2 2  180  4  180 

and hence cos 61  0.48481.

To estimate the accuracy of this approximation, we consider


3
f 3 (t x ) 
r2 ( x)  x  
3!  3

  
where tx is between and  . We then obtain
3 3 180

54
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

 
sin t x    1  
3 3

r2 (  )       0.000001.
3 180 3!  180  3!  180 

Therefore cos 61  0.48481 accurate to five decimal places.

Class Work
Approximate the number to four decimal places
a. sin 89  b. 4.03

2.3.5 Algebraic Operations on Series

Power series can be combined through the operations of addition, multiplication, and
division. The procedures for power series are similar to the way in which two polynomials are
added, multiplied, and divided that is, we add coefficients of like powers of x, use the
distributive law and collect like terms, and perform long division.
Problems involving multiplication or division of power series can be done with minimal fuss
using a computer algebra system. Combination two or more summations as a single summation
often requires a re-indexing, that is a shift in the index of summation.
Example 1 Adding Two Power Series
 
Write  n(n  1)cn x n2   cn x n1 as one power series.
n2 n 0

Solution: In order to add the two series, it is necessary that both summation indices start with
the same number and that the powers of x in each series be “in phase,” that is, if one series starts
with the same power. Note that in the given problem, the first series starts with x0, whereas the
second series starts with x1. By writing the first term of the first series outside of the summation
notation,
   

 n(n  1)cn x n2   cn x n1 =2 1c2 x 0 +  n(n  1)cn x n2   cn x n1

n2 n 0 n 3 n 0

we see that both series on the right side start with the same power of x, namely, x1. Now to got
the same summation index we are inspired by the exponents of x; we let k = n-2 in the first series
and at the sane time let k =n+1 in the second series. The right side becomes
 
2c 2   (k  2)( k  1)c k 1 x k   c k 1 x k .
k 1 k 1

55
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Remember, the summation index is a “dummy” variable; the fact that k = n-1 in one case and k=
n+1 in the other should cause no confusion if you keep in mind that it is the value of the
summation index that is important. In both cases k takes the values k=1, 2, 3,... for k = n+1. We
are now in a position to add series in (3) term by term:
  

 n(n  1)cn x n2   cn x n1 = 2c2   [(n  2)(n  1)cn1  cn1 ]x n .

n2 n 0 n 1

Example 2 Find the first three nonzero terms in the power series for a)
e x sin x b) tan x
Solution:
a) Using the power series for e x and sin x , we have
 x2 x3  x3 
e x sin x  1  x    ... x   ...
 2! 3!  3! 
We multiply these expressions, collection like terms just as for polynomials:
1  x  12 x 2  16 x 3  ...
x  16 x 3  ...

x  x 2  12 x 3  16 x 4  ...
 16 x 3  16 x 4  ...

x  x 2  13 x 3  ...

Thus e x sin x  x  x 2  13 x 3  ...

b) Using the power series of sin x and cos x , we have

x3 x5
x
  ...
sin x 3! 5!
tan x  
cos x x2 x4
1   ...
2! 4!
We use a procedure like long division:

56
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

x  13 x 3  52 x 5  ...
1  12 x 2  1
24 x 4  ... x  16 x 3  120
1
x 5  ...
x  13 x 3  1
24 x 5  ...
1
3 x 3  301 x 5  ...
1
3 x 3  16 x 5  ...
2
15 x 5  ...

Thus tan x  x  13 x 3  52 x 5  ...

Class Work: 1. Rewrite the given expression as a single power series.

 
a.  2nc
n 1
n x n 1   6c n x n 1
n 0

  
b.  n(n  1)cn x n  2 n(n  1)cn x n2  3 ncn x n
n2 n2 n 1

3. Use multiplication of division of power series to find the first three nonzero terms in the
power series for each function.
ln( 1  x)
a. y  e  x cos x
2
b. y  c. y  e x ln( 1  x)
ex

2.3.6 The Binomial Series

We have already acquainted with the Binomial Theorem, which states that if a and b are any real
numbers and k is a positive integer, then
k (k  1) k  2 2 k (k  1)( k  2) k 3 3
(a  b) k  a k  ka k 1b  a b  a b
2! 3!
k (k  1)( k  2)...( k  n  1) k  n n
 ...  a b  ...  b k .
n!
The traditional notation for the binomial coefficients is
k  k  k (k  1)( k  2)...( k  n  1)
   1    for n  1.
0 n n!
which enables us to write the Binomial Theorem in the abbreviated form
k
k
(a  b) k    a k  n b n
n 0  n 

57
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

In particular, if we put a  1 and b  x, we get

k
k
(1  x) k    x n . (1)
n 0  n 

In the case in which k is no longer a positive integer the expression for (1  x) k is no longer a
finite sum; it becomes an infinite series. To find this series we compute the Taylor series of
(1  x) k in the usual way:

f ( x)  (1  x) k f (0)  1
k 1
f ' ( x)  k (1  x) f ' (0)  k
f ( x)  k (k  1)(1  x) k 2
f " (0)  k (k  1)
f ( x)  k (k  1)( k  2)(1  x) k 3
f (0)  k (k  1)( k  2)
 
f ( n ) ( x)  k (k  1)...( k  n  1)(1  x) k  n f ( n ) (0)  k (k  1)...( k  n  1)

Therefore, the Taylor series of f ( x)  (1  x) k is


f ( n ) (0) n  k (k  1)( k  2)...( k  n  1) n

n 0 n!
x 
n 0 n!
x .

This series is called the binomial series. If its nth term is an, then we can show that

an1 x n1 k n
lim  lim x  x.
n a x n n n  1
n

Thus, by the Ratio Test, the binomial series converges if x  1 and diverges if x  1 . Even if it

is not easy to show the following theorem its result is very important.
Theorem 3.10

  1  x  1 if k  1
  1  x  1 if  1  k  0

k 
(1  x) k    x n for 
n 0  n   1  x  1 if k  0 and k is not an int eger
 all x if k is nonnegative int eger

Example 1 Find a power series representation for 1  x .

Solution: Using theorem 3.10 with k  12 , we obtain

1 (  1) 2 12 ( 12  1)( 12  2) 3
1 1
(  1)...( 12  n  1) n
1 1
1 x  1 x 2 2 x  x  ...  2 2 x  ...
2 2! 3! n!

58
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

which may be written

1 1 3 1  3  5    (2n  3) n
1 x  1 x  2 x 2  3 x 3  ...  (1) n 1 x  
2 2 2! 2 3! 2 n n!
for x  1 . The formula for the nth term of this series is valid provided n  2.

1
Example 2 Find the power series representation for f ( x)  and its radius of
4 x
convergence.
Solution: As given, f(x) in not in the form (1  x) k , so we rewrite it as follows:
1 / 2
1 1 1 x
  1  
4 x  x 2 4
41  
 4
Using the binomial series with k   12 and with x replaced by –x/4, we have
1 / 2
1    1  x 
n
1 1 x
 1      2   
4 x 2 4 2 n 0  n  4 
1   1  x   12  32   x   12  32 ...( 12  n  1)  x  n 
2

 1            ...      ...
2   2  4  2!  4 n!  4 
1 1 3 1 3  5 3 1  3  5  ...  (2n  1) n 
 1  18 x  2 x 2  x  ...  x  ...
2 2!8 3!8 3
n!8 n

We know from theorem 3.10 that this series converges when  x / 4  1, that is, x  4, so the

radius of convergence is R = 4.
Class Work 1
1. Use the binomial series to expand the given function as a power series. State the radius of
convergence.
1 1
a. 4
1 x4 b. c.
(1  x ) 3 2 x

2. a) Use the binomial series to expand 1 / 1  x 2 .

b) Use part (a) to find the power series for sin 1 x.
3.Approximate the number with an error less than 0.001.
a. 1.05 b. 3
9 c. 29 2 / 3

59
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

Exercises 2.3:

1. Find the interval of convergence and radius convergence of each of the following:
 
n! (n!) 2
i.  n x n ii.  n
xn
n 1 n n 0 ( 2n)!10

1  x  2
 n
xn
iii.  (1)
n 1
n 1

n
iv.   
n 1 2 n  1  x  1 
 
n
v. n
n 0
2
xn vi.  n x n
n 0 10
  n
1 2 n 1 n n
vii.
n 0 ( 4)
n
x viii. n 1
( 1)
n 1
( x  3) n

1  3  5    (2n  1) n 
sin x
ix.  x x.  2
n 1 3  6  9    (3n) n 1 n
2. Find the radius of convergence of the power series for positive integers c and d.


(n  c)! n (cn)!
a.  x b. 
n  0 n!( n  d )!
n n
n 0 ( n!) x
x
3. Find a power series representation for f ( x), f ' ( x), and  0
f (t )dt.
1 2 1 3
a. f ( x)  , x  b. , x 
2  7x 7 3  2x 2
4. Find a power series in x that has the given sum, and specify the radius of convergence.
x2 x3  t5
a. b. c.
1 x2 4  x3 (1  t 4 ) 2
5. Find the power series representation of
x x
a. cosh ii. tanh
2 3
6. Find the Taylor series expansion of each of the following series at the given point
x 
i. ,c0 ii. cos x, c 
1 x 2
2 1
iii. (1  x ) , c  0 iv. ln(3x), c=1

7. Approximate the following using the first 5 terms:

1

 7 
2

a. cos   x cos( x
3
b. )dx
 12  0
0.5 0.5

x e  sin( x
2  x2 2
c. dx. d. )dx
0 0

1 1
d. Approximate by P4(x) and R3(x) at c=1. Also approximate by means of
e x
e1.1
P4(x).

60
Calculus II Unit 4: Sequences and Series; Power Series by: Tadesse Bekeshie (PhD)

8. Find a power series representation for the expressions, and state the radius of convergence.
a. x(1  2 x) 2 b. 3 8  x c. (4  x) 3 / 2
9. Obtain a power series representation for f(x) by using the given relationship given and find
the radius of convergence.
x
a) f ( x)  sin 1 x; sin 1 x   (1 / 1  t 2 )dt
0
x
b) f ( x)  sinh 1
x; sinh 1
x   (1 / 1  t 2 )dt
0

61