APPM 4360/5360 Homework Assignment #4 Solutions Spring 2018

H
Problem #1 (20 points): Evaluate the integral C f (z)d z where C is the unit circle enclosing the origin and f (z)
is given by
p
(a) z + z 0 , |z 0 | > 1

(b) z/(4z 2 + 1)

Solution:
p p
(a) z + z 0 , |z 0 | > 1 Consider an analytic branch of z + z 0 such that branch cut joining z 0 and ∞ does not
p
cross the unit circle centered at z = 0. Then z + z 0 is analytic inside C and, by Cauchy theorem,
H p
C z + z 0 d z = 0.

(b) z/(4z 2 + 1),

z 1 1
= + ,
4z 2 + 1 4(2z − i ) 4(2z + i )
z = ±i /2 are the singularities of f (z) inside the contour. For each summand, we find

1 1
I I
d z = 2πi /8 = i π/4, d z = 2πi /8 = i π/4,
C 4(2z − i ) C 4(2z + i )
H
so C f (z)d z = i π/4 + i π/4 = i π/2.

H
Problem #2 (20 points): Evaluate the integral C f (z) d z where C is the unit circle centered at the origin for the
following f (z):
ez
(a)
z2
sin z
(b)
z4

Solution: Here the only singular point inside the unit circle is z = 0. We expand the numerators in Taylor series
around zero and use the integration of powers formula:

(a)
ez 1 X ∞ zn ∞
X zk
= = ,
z 2 z 2 n=0 n! k=−2 (k + 2)!

power z −1 corresponds to k = −1, thus C f (z) d z = 2πi .

H

(b)
sin z 1 1 z
= 3− + +··· ,
z4 z 6z 120
H
so C f (z) d z = −2πi /6 = −i π/3.

H
Problem #3 (20 points): Evaluate the integrals C f (z)d z over a contour C , where C is the boundary of a square
with diagonal opposite corners at z = −(1 + i )R and z = (1 + i )R, where R > a > 0, and where f (z) is given by the
following (use Eq. (1.2.19) as necessary):
sin z
(a)
z2
cosh z
(b)
z

1
Solution:
sin z
(a) ,
z2
∞ 2n+1
sin z 1 X n z
= (−1) =
z2 z 2 n=0 (2n + 1)!
z3
µ ¶
1 1 z
= 2 z− +... = − +...,
z 6 z 6
and the only singular point z = 0 is inside the contour. Deforming the contour to a circle around z = 0 and
using Cauchy theorem, we find
sin z 1
I I
2
d z = d z = 2πi .
C z C z

cosh z
(b) ,
z
cosh z 1 X ∞ z 2n
= =
z z n=0 (2n)!
z2
µ ¶
1 1 z
= 1+ +... = + +...,
z 2 z 2
and the only singular point z = 0 is inside the contour. Deforming the contour to a circle around z = 0 and
using Cauchy theorem, we find
cosh z 1
I I
dz = d z = 2πi .
C z C z

Problem #4 (25 points): Let f (z) be an entire function, with | f (z)| ≤ C |z| for all z, where C is a constant. Show
that f (z) = Az, where A is a constant.

Solution: Ussing the (generalized) Cauchy formula,

1 f (ζ)
I
f ′ (z) = d ζ,
2πi C (ζ − z)2

where C = {|ζ − z| = R} is the circle of radius R around z in ζ-plane. Then

1 | f (ζ)|
I
′
| f (z)| ≤ |d ζ| ≤
2π C |ζ − z|2
Z2π
1 C (|z| + R)
≤ Rd θ = C (1 + |z|/R) →R→∞ C ,
2π 0 R2
so f (z) is entire and bounded, so it is constant by Liouville theorem. Let f ′ (z) = A, then f (z) = Az + B , where A,
′

B are constants. But, since | f (z)| ≤ C |z| for all z, taking |z| → 0, we get B = 0. Thus, f (z) = Az as claimed.

Problem #5 (20 points): Discuss whether the sequence {1/(nz)4 }∞ 1 converges and whether the convergence is
uniform for: 0 < α < |z| < 1. Discuss whether the convergence is uniform if α = 0.

Solution:
1 1 1
lim 4
= 4 lim 4 = 0,
n→∞ (nz) z n→∞ n
so the sequence converges pointwise for every z. If |z| > α > 0, then
¯ ¯
¯ 1 ¯ 1 1
¯ (nz)4 ¯ = n 4 |z|4 ≤ α4 n 4 ,
¯ ¯

2
which is a convergent numerical sequence. Thus, the convergence is uniform for 0 < α < |z|. However, for α = 0
convergence is not uniform since 1/|z|4 is unbounded above in this case.

P∞ n
Problem #6 (20 points): Show that the following series converges uniformly in the given region: n=1 z ,
0 ≤ |z| < R, R < 1

Solution: ¯∞ ¯ ∞
¯ X n¯ X ∞ R
n
Rn =
X
¯ z ¯≤ |z| ≤ ,
¯
n=1
¯
n=1 n=1 1−R
i.e. the series is bounded above by a convergent numerical series which means uniform convergence by
Weierstrass M-test.

P∞
Problem #7 (20 points): Find the radius of convergence of the series 0 a n (z) where a n (z) is given by:

(a) z 3n

(b) n n z 2n

Solution: We apply the ratio test.

(a) z 3n = (z 3 )n , ¯ ¯ ¯ 3n ¯
¯ an ¯ ¯ z ¯ 1
¯ ¯ z 3(n+1) ¯ = |z|3 ,
¯ ¯=¯ ¯
¯a
n+1
therefore the series converges for |z| < 1 and radius of convergence R = 1.

(b) n n z 2n
n n z 2n
¯ ¯ ¯ ¯
¯ an ¯ ¯ ¯
¯=¯
¯ ¯ (n + 1)( n + 1)z 2(n+1) ¯ =
¯ ¯
¯a
n+1
1
= →n→∞ 0,
(n + 1)(1 + 1/n)n |z|2
which shows that R = 0 (series converges only for z = 0).

Problem #8 (15 points): Find Taylor series expansions around z = 0 of the following functions in the given
regions:
1
(a) 1−z 2
, |z| < 1

(b) cosh z, |z| < ∞

cos z−1
(c) z2
, 0 < |z| < ∞

Solution:
1
(a) 1−z 2
, |z| < 1
1 ∞ ∞
2 n
z 2n .
X X
= (z ) =
1 − z 2 n=0 n=0

(b) cosh z, |z| < ∞

e z + e −z X ∞ z 2n
cosh z = = .
2 n=0 (2n)!

3
 cos z−1
(c) z2
, 0 < |z| < ∞
(−1)n z 2n
µ∞ ¶
cos z − 1 1 X
= − 1 =
z2 z 2 n=0 (2n)!
X∞ (−1)n z 2n−2 X∞ (−1)n+1 z 2n
= = .
n=1 (2n)! n=0 (2(n + 1))!

Problem #9 (20 points): Use the Taylor series for (1 + z)−1 about z = 0 to find the Taylor series of log(1 + z) about
z = 0 for |z| < 1.

Solution: The Taylor series for (1 + z)−1 is just the geometric series

1 ∞
(−1)n z n ,
X
=
1 + z n=0

and we know that it converges uniformly in |z| < 1. Since (1 + z)−1 d z = log(1 + z) + c and since the above series
R

converges uniformly so we can integrate it termwise, we find that

∞ z n+1 X∞ zn
(−1)n (−1)n−1 .
X
log(1 + z) = =
n=0 n + 1 n=1 n

Problem #10 (20 points): Use the Taylor series representation of (1 + z)−1 around z = 0 for |z| < 1 to find a series
representation of 1/(1 + z) for |z| > 1. (Hint: use 1/(1 + z) = 1/(z(1 + 1/z))

Solution: The Taylor series for (1 + z)−1 around z = 0 is just the geometric series

1 ∞
(−1)n z n ,
X
=
1 + z n=0

and we know that it converges in |z| < 1. For |z| > 1, 1/|z| < 1, so we have

1 1 ∞ (−1)n
1 X X∞ (−1)n
= = = .
1 + z z(1 + 1/z) z n=0 z n n=0 z
n+1

Extra-Credit Problem #11 (5 points): Use the binomial expansion and Cauchy integral formula to
evaluate
1 2n d z
I µ ¶
z+
|z|=1 z z
Recall the binomial expansion à !
n n
n
a k b n−k
X
(a + b) =
k=0 k
where à !
n n!
= .
k k!(n − k)!
Use this result to establish the following real integral formula:
Z2π
1 (2n)!
(cos θ)2n d θ = .
2π 0 4n (n!)2

4
Solution: Using the binomial expansion, we get
¶2n à !
µ
1 2n 2n
z k (1/z)2n−k .
X
z+ =
z k=0 k

Substituting this into the integral we find

¶2n à !
I µ
1 dz X 2n 2n I dz
z+ = z k (1/z)2n−k =
|z|=1 z z k=0 k |z|=1 z
à !I à !
2n dz 2n (2n)!
= = 2πi = 2πi ,
n |z|=1 z n (n!)2

since only z −1 term contributes which corresponds to k = 2n − k or k = n. For the real integral we find

Z2π Ã !2n
e i θ + e −i θ i eiθd θ
Z2π
1 2n1
(cos θ) d θ = =
2π 0 2πi 0 2 eiθ
¢2n
z + 1z
¡
1 dz (2n)!
I
= = n ,
2πi |z|=1 22n z 4 (n!)2
where we used the result for the integral in the last equality.