Sequences and series

Power series
Taylor and Maclaurin series

BMEG2410: Complex Analysis Part III

Prof. Scott YUAN

Department of Biomedical Engineering
The Chinese University of Hong Kong

September 2020
Lecture notes: Courtesy of Prof. Hongsheng LI
 Sequences and series
Power series
Taylor and Maclaurin series

Outline

1 Sequences and series

2 Power series

3 Taylor and Maclaurin series

 Sequences and series
Power series
Taylor and Maclaurin series

Reading materials:
Power series, Taylor series: Kreyszig 15.1, 15.2, 15.3, 15.4
 Sequences and series
Power series
Taylor and Maclaurin series

1 Sequences and series

2 Power series

3 Taylor and Maclaurin series

 Sequences and series
Power series
Taylor and Maclaurin series

Sequences
A sequence is an indexed collection of (real or complex) numbers:
zn , for n = 1, 2, . . . , ∞ or z1 , z 2 , z 3 , · · · or simply {zn }
Convergence of sequences
A sequence is convergent iff there exists a constant c such that
lim zn = c
n→∞
A divergent sequence is one that does not converge

Example: sequence {in /n} = {i, −1/2, −i/3, 1/4, · · · } is convergent → 0

Example: sequence {in } = {i, −1, −i, 1, · · · } is divergent
Convergence of complex number sequences
A sequence z1 , z2 , · · · , zn , · · · of complex numbers zn = xn + iyn converges to
a + ib iff the sequence of real parts x1 , x2 , · · · converges to a, and the
sequence of imaginary parts y1 , y2 , · · · converges to b

Example: sequence {zn } with zn = xn + iyn = 1 − 1/n2 + i(2 + 4/n)

converges to 1 + 2i, because real parts {1 − 1/n2 } converge to 1, and
imagenary parts {2 + 4/n} converge to 2
 Sequences and series
Power series
Taylor and Maclaurin series

Series
Given a sequence z1 , z2 , · · · , zn , · · · , the series could be formed as the
summations of terms z1 , z2 , · · · , zn , · · · ∞
X
s1 = z 1 , s2 = z1 + z2 , ··· , sn = zn = z1 + z2 + · · · + zn
n=1

Convergence of series
A series is convergent iff there exists a constant s such that
lim sn = s
n→∞
A series is divergent if it is not convergent
A series sn = z1 + z2 + · · · is absolutely convergent if the sum of the
absolute values of the terms is convergent
X∞
|zn | = |z1 | + |z2 | + |z3 | + · · ·
n=1
If a series is absolute convergent, it is convergent
If z1 + z2 + · · · converges but |z1 | + |z2 | + · · · diverges, the series
z1 + z2 + · · · is called conditionally convergent
Example: the series 1 − 21 + 13 − 14 + · · · converges, but 1 + 1
2
+ 1
3
+ 1
4
+ ···
does not. It is conditionally convergent.
 Sequences and series
Power series
Taylor and Maclaurin series

Sequences and series

Geometric series X
The geometric series is defined by the terms zn = q n where q ∈ C. The sum of
this sequence is
X n 1
convergent, when |q| < 1. In that case q = ;
1−q
n≥0
divergent, when |q| ≥ 1.

Examples:
zn = 1/n, n = 1, 2, ..., ∞: convergent sequence but divergent series
zn = 1/n2 , n = 1, 2, ..., ∞: convergent sequence and series
zn = (−1)n /n, n = 1, 2, ..., ∞: convergent sequence and series
zn = cos(n), n = 1, 2, ..., ∞: divergent sequence and series
 Sequences and series
Power series
Taylor and Maclaurin series

Convergence and divergence tests of series

Divergence
P
Given a series sn = n zn , if lim zn 6= 0, the series is divergent
n→∞

Cauchy’s convergence principle

P
A series sn = n→∞ zn is convergent iff for every  > 0 (no mattere how
small), there exist an N (usually dependent on ) such that
|zn+1 + zn+2 + · · · + zn+p | ≤  for all n > N and

Comparison test
X X
A series zn is absolute convergent if there is a convergent series bn
n→∞ n→∞
such that |z1 | ≤ b1 , |z2 | ≤ b2 , · · ·
 Sequences and series
Power series
Taylor and Maclaurin series

Convergence and divergence tests of series

Ratio test 1
P
A series n→∞ zn is absolutely convergent, if there exists a constant q < 1
such that  
 zn+1 
 zn  ≤ q < 1 for all n greater than some N
 

The series is divergent, if

 
 zn+1 
 zn  ≥ 1 for all n greater than some N
 

Ratio test 2
 
 zn+1 
If a series with terms zn , where zn 6= 0, such that  =L
zn 
If L < 1, the series converges absolutely
If L > 1, the series diverges
If L = 1, the series may converge or diverge (inconclusive)
 Sequences and series
Power series
Taylor and Maclaurin series

Convergence and divergence tests of series

Root test 1
A series with terms zn is absolutely convergent if there exist some constant
q<1 pn
|zn | ≤ q < 1 for all n greater than some N
The series is divergent, if
p
n
|zn | ≥ 1 for infinitely many n

Root test 2
p
n
A series with terms zn is such that lim |zn | = L, then
n→∞
If L < 1, the series converges absolutely
If L > 1, the series diverges
If L = 1, the series may converge or diverge (inconclusive)
 Sequences and series
Power series
Taylor and Maclaurin series

Examples
1
The series with term zn = n(n−1)
, n ≥ 2 is convergent because:
(1) We can apply Cauchy’s convergence principle, Assume we have some  > 0,
|zn+1 + zn+2 + zn+3 + · · · zn+p |
1 1 1 1 1 1
=( − )+( − ) + ···( − )
n n+1 n+1 n+2 n+p−1 n+p
1 1
= −
n n+p
The largest value for this equation is obtained when p → ∞, and the value
become n1 . We could determine N as any value N > 1 .
(2) We apply ratio test 1,  
  1   
 zn+1   (n+1)n   n − 1   2 
 
 = =
 zn   1   n + 1   = 1 − <1
n + 1

n(n−1)

Cannot determine convergence use ratio test 1

(3) We apply ratio test 2, 
   1     
 = lim  (n+1)n  = lim  n − 1  = lim 1 − 2  = 1
 zn+1     
lim  1
n→∞ zn  n→∞  n(n−1)  n→∞  n + 1  n→∞  n + 1
Cannot determine convergence or divergence with ratio test 2
 Sequences and series
Power series
Taylor and Maclaurin series

Examples

1
The series with term zn = , n ≥ 1 is convergent because
n2
(1) We apply comparison test
1
|zn | ≤ bn = for n ≥ 2,
n(n − 1)
where bn is a convergent series

(2) We apply ratio test 2,

2
     
 zn+1 
 = lim  n  = lim 1 − 2n + 1  = 1
   
lim 
n→∞ zn  n→∞  (n + 1)2  n→∞  (n + 1)2 
Cannot determine convergence or divergence
n
The series with term zn = (−1) n!
, n ≥ 1 is convergent because
We apply ratio test 2  
 zn+1 
lim  = 1 =0
n→∞ zn  n+1
 Sequences and series
Power series
Taylor and Maclaurin series

Examples

2
The series with term zn = e−n , n ≥ 1 is convergent because
We apply root test 2
lim |zn |1/n = lim e−n = 0
n→∞ n→∞
 Sequences and series
Power series
Taylor and Maclaurin series

1 Sequences and series

2 Power series

3 Taylor and Maclaurin series

 Sequences and series
Power series
Taylor and Maclaurin series

Definition

A power series is a series with term an (z − z0 )n where an is a sequence of

complex numbers and z, z0 are a couple of complex numbers
X
sn = an (z − z0 )n
n≥0

Power series are used to define a complex function f (z) with a variable z
X
f (z) = an (z − z0 )n
n≥0

z0 is called the center of the power series

When given fixed {an } and z0 , we are interested to study the convergence
of the power series with different variable value z
 Sequences and series
Power series
Taylor and Maclaurin series

Convergence of a power series

Convergence of a power series

For a power series with term an (z − z0 )n
The power series converges at center z0 (when z = z0 for f (z))
If the power series converges for z = z1 , then converges absolutely in the
open disk centered at z0 and of radius |z1 − z0 | (excluding z1 )
If a power series diverges for z = z2 , then it diverges everywhere away
from the closed disk centered at z0 and of radius |z2 − z0 |

NOTE: The natural shape of a convergence domain is a disk centered at z0 ;

hence we will always use the notion of convergence radius
 Sequences and series
Power series
Taylor and Maclaurin series

Convergence radius of power series

Cauchy-Hadamard formula
 
a
Given a power series with term an (z − z0 )n , we create a sequence  n+1

an ,
n = 1, 2, · · ·
 
 an+1 
If lim   = 0, then convergence radius R = ∞
n→∞ an 
 
 an+1 
If lim   = L∗ , where L∗ 6= 0 or L∗ 6= ∞, then R = 1
n→∞ an  L∗
 
 an+1 
If lim   = ∞, the convergence radius R = 0
n→∞ an 

NOTE: When a power series can be expressed as a sum of independent power

series with different convergence radii (but same center), the convergence
radius of the sum is the smallest of the independent convergence radii
 Sequences and series
Power series
Taylor and Maclaurin series

Example

∞
X (2n)!
Find the convergence radius of the power series (z − 3i)n
n=0
(n!)2

(2n)!
Solution: we can find that an = and z0 = 3i
(n!)2

(n!)2
      
 an+1  (2n + 2)! (2n)! (2n + 2)!
lim   = lim = lim ·
n→∞ an  n→∞ ((n + 1)!)2 (n!)2 n→∞ (2n)! ((n + 1)!)2
(2n + 2)(2n + 1)
= lim
n→∞ (n + 1)2
=4

1
Therefore, the convergence radius R =
4
 Sequences and series
Power series
Taylor and Maclaurin series

Exercises

Compute the convergence radius of the power series with coefficients

an = 2 n , n ≥ 0
an = n 2 2 n , n ≥ 0
an = 2 n + 3 n , n ≥ 0
an = 1/n2 , n ≥ 1
an = 1/n!, n ≥ 0
2
an = e−n , n ≥ 0
2
an = e n , n ≥ 0
 Sequences and series
Power series
Taylor and Maclaurin series

Functions given by power series

Analytic power series X

The power series n≥0 an (z − z0 )n with convergence radius R > 0 defines
P
a function which is analytic in the open disk of radius R centered at z0 .
If another sequence bn of the power series defines the same function, then
bn = an . In other words, if a function f (z) can be represented by a power
series with any center z0 , the representation is unique

Derivative and integral of a power series X

The derivative of a power series with coefficients an , n ≥ 0 is given by
X
f 0 (z) = (n + 1)an+1 (z − z0 )n
n≥0
with the same convergence radius as the initial power series. The integral of
the same power series is given by (unchanged convergence radius)
Z X an−1
F (z) = f (z) dz = Constant + (z − z0 )n
n
n≥1

NOTE: Natural extension to higher order derivatives.

 Sequences and series
Power series
Taylor and Maclaurin series

1 Sequences and series

2 Power series

3 Taylor and Maclaurin series

 Sequences and series
Power series
Taylor and Maclaurin series

Series from analytic functions

Taylor series X
Consider a function analytic in an open domain D and a point z0 ∈ D. Then,
for any z in a disk centered at z0 contained in D
X f (n) (z0 ) 1
f (z) = (z − z0 )n where an = f (n) (z0 )
n! n!
n≥0
or by Cauchy’s integral formula
f (z ∗ )
I
1
an = ∗
dz ∗
2πi C (z − z0 )n+1
where C is a simple couterclockwise closed path enclosing z0 and f (z) is
analytic in the domain containing C and all points inside.
Taylor series are power series. The convergence of the Taylor series is absolute.

This theorem shows that analytic functions are similar to polynomials (of
infinite degree) within disks, and that all the information needed to describe
them is how they behave in the neighborhood of disk centers.
Maclaurin expansion is a Taylor expansion for z0 = 0.
 Sequences and series
Power series
Taylor and Maclaurin series

Series from analytic functions

The proof of Taylor expansion uses Cauchy’s integral formula expressed as

f (z ∗ ) ∗ f (z ∗ )
I I
1 1 1
f (z) = dz = dz ∗
∗
2iπ C z − z 2iπ C z ∗ − z0 1 − z − z0
z ∗ − z0 n
f (z ∗ ) X z − z0
I 
1
= dz ∗
2iπ C z ∗ − z0 z ∗ − z0
n≥0
f (z ∗ )
I
X 1
= (z − z0 )n ∗
dz ∗
2iπ C (z − z0 )n+1
n≥0 | {z }
f (n) (z0 )/n!
 
where the condition on z ensures that  zz−z 0 
0 −z  < 1.

0
 Sequences and series
Power series
Taylor and Maclaurin series

Coefficients of Taylor series

Depending on the available properties of the analytic function f (z), various

strategies are available to find the Taylor series coefficients an .

Finding Taylor series coefficients an

If all the derivatives are easy to compute, use Taylor expansion. Example:
ez , cos z, ln(1 + z), (1 + z)α , etc.;
If the function can be expressed as a sum of elementary known functions,
then the power series coefficients are the sums of the individual power
series coefficients;
If the function results from a “simple” change of variables in a known
functions (e.g., z → z 2 ), make this change of variables in the power series
directly;
If the function satisfies a specific equation (e.g., differential equation,
inverse of some function, etc.), use it to obtain an induction formula for
the power series coefficients.
 Sequences and series
Power series
Taylor and Maclaurin series

Series from simple functions

Exponential functionX
X zn
ez = , with convergence in C
n!
n≥0

Proof: since f (z) = ez then f (n) (z) = f (z). Hence, according to Taylor’s
formula f (0) 1
an = =
n! n!
8
exp(x)
7 Taylor 2 terms
Taylor 3 terms
6 Taylor 4 terms
Taylor 5 terms
5

0
−1 −0.5 0 0.5 1 1.5 2
x
 Sequences and series
Power series
Taylor and Maclaurin series

Series from simple functions

Trigonometric functions X
X (−1)n z 2n 
cos z = 

(2n)! 

n≥0
X (−1)n z 2n+1 with convergence in C
sin z =



(2n + 1)! 
n≥0

eiz + e−iz eiz − e−iz

Proof: from cos z = and sin z = .
2 2i
1 3
sin(x)
0.5 2.5 Taylor 2 terms
0 2 Taylor 3 terms
Taylor 4 terms
−0.5 1.5

−1 1

−1.5 0.5

−2 0
cos(x)
−2.5 −0.5
Taylor 2 terms
−3 Taylor 3 terms −1
Taylor 4 terms
−3.5 −1.5
−1 0 1 2 3 −1 0 1 2 3
x x
 Sequences and series
Power series
Taylor and Maclaurin series

Series from simple functions

Geometric series X
1 X n
= z , with convergence in the unit disk centered at (0, 0)
1−z
n≥0

1/(1−x)
3 Taylor 2 terms
Taylor 3 terms
2.5 Taylor 4 terms
Taylor 5 terms

1.5

0.5

−0.6 −0.4 −0.2 0 0.2 0.4 0.6

x
 Sequences and series
Power series
Taylor and Maclaurin series

Series from simple functions

Logarithm X
X zn
ln(1 − z) = − , with convergence in the unit disk centered at (0, 0)
n
n≥1

Proof: By integrating the geometric series.

0.6 ln(1−x)
Taylor 2 terms
0.4 Taylor 3 terms
0.2 Taylor 4 terms
Taylor 5 terms
0

−0.2

−0.4

−0.6

−0.8

−1

−1.2
−0.6 −0.4 −0.2 0 0.2 0.4 0.6
x
 Sequences and series
Power series
Taylor and Maclaurin series

Other examples

Function to develop in power series

1
f (z) =
(z − 1)(z − 2)

Steps of the solution

1 1 1/2
1 Partial fractions decomposition: = −
(z − 1)(z − 2) 1−z 1 − z/2
2 Series coefficients of 1/(1 − z): bn = 1, n ≥ 0
3 Series coefficients of 1/(1 − z/2): cn = 2−n , n ≥ 0

Result
Hence, the series coefficients of f (z) are
an = bn − cn /2 = 1 − 2−n−1
Convergence in the open disk |z| < 1, where f (z) is analytic.
 Sequences and series
Power series
Taylor and Maclaurin series

Other examples

Function to develop in power series

2
f (z) = e−z

Steps of the solution

X zn
1 exponential series: ez =
n≥0 n!
2 X (−z 2 )n
2 replace z by −z 2 : e−z =
n≥0 n!

Result
Hence, the series coefficients of f (z) are
 n/2
 (−1) , if n is even;
an = (n/2)!
0, if n is odd.

Convergence in C, where f (z) is analytic.
 Sequences and series
Power series
Taylor and Maclaurin series

Other examples

Function to develop in power series

f (z) = (1 + z)α

Steps of the solution

1 specific property of f (z): αf (z) = (1 + z)f 0 (z)
2 replace f (z) and f 0 (z) by their series:
X X
α an z n = (1 + z) (n + 1)an+1 z n
n≥0 n≥0

3 identify coefficients of same power: αan = (n + 1)an+1 + nan

Result
Hence, the series coefficients of f (z) are
α−n+1 (α − n + 1)(α − n + 2) . . . α
an = an−1 =
n n!
Convergence in the open disk |z| < 1, where f (z) is analytic.