Professional Documents
Culture Documents
Solution: Let C1 : |z| = 1 and C2 : |z − 1| = 1. Let the common region between them is Ω
C1 C2
Ω
1+ω −ω 2 ω 1
f1 (1) = = =ω f1 (0) = = = ω̄ = ω 2
1 + ω2 −ω ω2 ω
Hence f1 (C1 ) = line joining 0 and ω and f1 (C2 ) = line joining 0 and ω 2
C2
C1
2π
Now we rotate the region Ω by 2π
3 clockwise by the conformal map f2 (z) = e−i 3 z = ωz
C1 Ω
C2
3
Now we map the common region Ω to the upper half of the plane by the conformal map f3 (z) = z 2
1
Now we want to map the upper half plane to inside of the unit circle. We do it with the conformal map
z−ω
f4 (z) = z−ω 2
Hence the final conformal map which maps the region Ω to the inside of unit disk is
ω2z + 1
z+ω z+ω
f4 ◦ f3 ◦ f2 ◦ f1 (z) = f4 ◦ f3 ◦ f2 = f4 ◦ f3 ω = f4 ◦ f3
z + ω2 z + ω2 ωz + 1
h 2 i3
32 ! ω z+1 2
2
ω z+1 ωz+1 −ω
= f4 =h 3
ωz + 1 i
ω 2 z+1 2
ωz+1 − ω2
1
Solution: Let C1 : |z| = 1 and C2 : z − 2 = 12 . Let the common region between them is Ω
C1 C2
Ω
i i(−i − 1) 1 i −i i(1 − i) 1 i
f (i) = = = − f (−i) = = =+ +
i−1 2 2 2 −i − 1 2 2 2
Hence f (C1 ) = is the line parallel to imaginary axis passing through 21 . f (C2 ) = imaginary axis. Hence
the region Ω is mapped to the 21 width strip parallel to imaginary axis enclosed between imaginary axis
and ℜ(z) = 12 .
2
Ω
C2 C1
Now we rotate the region Ω by 90◦ counter clockwise with the conformal map f2 (z) = iz
C1
C2
Since the strip width is 21 we take the conformal map f3 (z) = (ez ) 2 = (ez )2π = e2πz which maps the strip
1
Hence the final conformal map which maps the region between the two circles C1 and C2 to the upper half
plane is
z z 2πiz
f3 ◦ f2 ◦ f1 (z) = f3 ◦ f2 = f3 i = e z−1
z−1 z−1
We map the semi arc containg i to the negative real axis then ultimately to the right half plane.
z+1
We first map 1 → ∞ and −1 → 0 with the conformal map f1 (z) = z−1 . Then
3
C
Now we will only concentrate on the lower ray i.e. {z | ℜ(z) = 0, ℑ(z) ≤ 0}. Now we want to rotate
this by 90◦ clockwise to map it to the real axis by the conformal map f2 (z) = −iz. Hence f2 (1) = −i.
√
Now we know f3 (z) = z maps the C\negative real axis to right half plane.
π π π
Hence f3 (−i) = e−i 4 . Hence at this stage ∞ is mapped to e−i 4 . Now we have to map e−i 4 to 1. This is
achieved by conformal map of the form f4 (z) = az + b where a ∈ R, b ∈ C, b = b1 + ib2 . Therefore
π π a a a
f4 e−i 4 = 1 ⇐⇒ ae−i 4 + b = 1 ⇐⇒ √ (1 − i) + b = ⇐⇒ b2 = √ , b1 = √ − 1
2 2 2
√ √
We can take a = 2, b1 = 0 and b2 = 1. then we have f4 (z) = 2z + i. Hence now ∞ is mapped to 1
Now we will rotate the right half plane by π2 counter clockwise to map it to upper half plane by the
conformal map f5 (z) = iz
z−i
Finally we use the map f6 (z) = z+i which maps the upper half plane to the inside of unit disk. And
also we have f6 (i) = ∞. Hence the final map
f6 ◦ f5 ◦ f4 ◦ f3 ◦ f2 ◦ f1 (z)
maps ∞ → ∞. and this map, maps the arc |z| = 1, y ≥ 0 on the outside of the unit circle so that the
points at ∞ correspond to each other
4
Problem 4 Ahlfors Page 108: Problem 1
Compute Z
xdz
γ
Solution:
γ is the directed line segment from 0 to 1 + i. Hence z = t(1 + i) where t ∈ [0, 1]. Then we have
dz = (1 + i)dt
then
Z Z 1
xdz = ℜ(t(1 + i))(1 + i)dt
γ 0
Z 1
= t(1 + i)dt
0
Z 1
= (1 + i) tdt
0
1
t2
= (1 + i)
2 0
1+i
=
2
Solution:
dz = ireiθ dθ
5
Z Z 2π
xdz = ℜ(reiθ ) ireiθ dθ
|z|=r 0
Z 2π
= ℜ(r (cos θ + i sin θ)) (ir (cos θ + i sin θ)) dθ
0
Z 2π
2
= ir cos θ (cos θ + i sin θ) dθ
0
Z 2π Z 2π
= ir2 cos2 θdθ + i cos θ sin θdθ
0 0
Z 2π
i 2π
Z
2 1
= ir (cos 2θ + 1)dθ + sin 2θdθ
2 0 2 0
Z 2π
2 1
= ir dθ
2 0
1
= ir2 (2π − 0)
2
= iπr2
•
r2
Z Z Z
1 1
xdz = (z + z̄)dz = z+ dz
|z|=r |z|=r 2 |z|=r 2 z
r2
Z Z
1 1
= zdz + dz
2 |z|=r 2 |z|=r z
| {z }
=0
As f is analytic
r2
= 2πi = iπr2
2
6
regions. The notion of winding number leads to a quick proof of one part of the theorem, namely
that the complement of a Jordan curve γ has at least two components. This will be so if there exists
a point a with n(γ, a) ̸= 0.
We may assume that ℜ(z) > 0 on γ, and that there are points z1 , z2 ∈ γ with ℑ(z1 ) <
0, ℑ(z)2 > 0. These points may be chosen so that there are no other points of γ on the line segments
from 0 to z1 and from 0 to z2 . Let γ1 and γ2 be the ares of γ from z1 to z2 (excluding the end points).
Let σ1 be the closed curve that consists of the line segment from 0 to z1 followed by γ1 and the
segment from z2 to 0 , and let σ2 be constructed in the same way with γ2 in the place of γ1 . Then
σ1 − σ2 = γ or −γ.
The positive real axis intersects both γ1 and γ2 (why?). Choose the notation so that the
intersection x2 farthest to the right is with γ2 (Figure). Prove the following:
(c) the first intersection x1 of the positive real axis with γ lies on γ1 ;
(e) there exists a segment of the positive real axis with one end point on γ1 , the other on γ2 , and
no other points on γ. The points x between the end points satisfy n(γ, x) = 1 or -1 .
Solution:
(a) We claim that x2 lies outside the curve σ1 . Indeed, consider the ray joining 0 and x2 . This ray can
not intersect γ beyond the point x2 because x2 is the farthest point. Thus the whole ray (x2 , ∞)
is contained in one region defined by the curve, but γ must be bounded. Thus x2 must lie in the
unbounded region defined by σ1 .
Now pick z ∈ γ2 . Then x2 , z are path connected, hence they must lie in the same connected
region of σ1 . Hence n (σ1 , z) = n (σ1 , x2 ) = 0
(b) Let S := {z | z ∈ γ, z ∈ R}. Since S is compact we can pick the smallest element of S, considered as
a subset of R, let’s call it x1 . We claim that n (σ1 , x) = n (σ2 , x) = 0 for x ∈ (0, x1 ). By Lemma 2 ,
it suffices to prove that γ1 , γ2 doesn’t pass through any point to the left of x1 , which is obvious by
the construction of x. Also, the path z1 → 0 → z2 intersects the X-axis only at 0 .
(c) Suppose the first intersection x1 of the positive real axis with γ does not lie on γ1 ; then it must lie
on γ2 . In part (b) we proved that n (σ1 , x) = 0 for x ∈ (0, x1 ). But now we can consider the path
from x to x1 along the positive X-axis and then use the path along γ2 from x1 to x2 . Thus x and x2
are path-connected and both are in the same region of σ1 . But then 1 = n (σ1 , x) = n (σ1 , x2 ) = 0,
where the first equality is from part (a) and the last equality is from part (b).
(d) Because x1 lies on γ1 , again by applying Lemma 2, we get n (σ2 , x1 ) = 1. Again, because any z ∈ γ2
is path-connected along γ2 , we must have n (σ2 , z) = 1
7
(e) Consider the sets S1 := {z ∈ R | z ∈ γ1 } and S2 := {z ∈ R | z ∈ γ2 }. Since S1 and S2 are compact
we can find two points a ∈ S1 , b ∈ S2 such that there is no point of γ between a and b, otherwise
we can construct a sequence of points in S1 or S2 such that it converges to a point in another set.
contradicting the fact that S1 , S2 are closed and disjoint. Now pick point x between a and b. Let us
assume γ = σ1 − σ2 . Then n(γ, x) = n (σ1 , x) − n (σ2 , x). But n (σ1 , x) = n (σ1 , b) = 0 by part (a)
and n (σ2 , x) = n (σ1 , a) = 1 by part (d).
eζ dζ
Z Z
1 f (ζ)dζ 1
f (z) = =
2πi |ζ|=1 ζ −z 2πi |ζ|=1 ζ −z
Hence
eζ ez
Z Z
0 1
1 = e = f (0) = dζ ⇐⇒ dz = 2πi
2πi |ζ|=1 ζ |z|=1 z
Solution:
(z + i) − (z − i)
Z Z
dz 1
2
= dz
|z|=2 z +1 2i |z|=2 (z + i)(z − i)
Z
1 1 1
= − dz
2i |z|=2 z − i z + i
"Z Z #
1 1 1
= dz − dz
2i |z|=2 z − i |z|=2 z + i
1
= [2πi − 2πi] = 0
2i
8
Problem 10 Ahlfors Page 120: Problem 3
Compute
|dz|
Z
|z|=ρ |z − a|2
̸ ρ. Hint: make use of the equations z z̄ = ρ2 and
under the condition |a| =
dz
|dz| = −iρ .
z
Solution:
• Let f (z)ez Then we have
(n − 1)! (n − 1)! eζ dζ
Z Z
f (ζ)dζ
ez = f ((n−1)) (z) = n
=
2πi |ζ|=1 (ζ − z) 2πi |ζ|=1 (ζ − z)n
Therefore
(n − 1)! ez ez
Z Z
0 2πi
f (0) = e = 1 = dz ⇐⇒ dz =
2πi |ζ|=1 zn |ζ|=1 z n (n − 1)!
9
• We have Z
1 f (ζ)dζ
f (z) =
2πi |ζ|=1 ζ −z
4 cases possible
Case 1: m ≥ 0, n ≥ 0
Case 2: m < 0, n ≥ 0
zn zn
Z Z Z
n m
z (1 − z) dz = dz = (−1)m dz
|z|=2 |z|=2 (1 − z)|m| |z|=2 (z − 1)|m|
2πi(−1)m d|m|−1 n 2πin!(−1)m
m n
= z = = 2πi(−1)
(|m| − 1)! dz |m|−1 (|m| − 1)!(n − (|m| − 1))! |m| − 1
z=1
If |m| − 1 > n then the above is zero
Case 3: m ≥ 0, n < 0
(1 − z)m d|n|−1
Z Z
2πi
z n (1 − z)m dz = dz = (1 − z)m
|z|=2 |z|=2 z |n| (|n| − 1)! dz |n|−1
z=0
2πim!(−1)|n|−1
|n|−1 m
= = 2πi(−1)
(|n| − 1)!(m − (|n| − 1))! |n| − 1
If |n| − 1 > m, we interpret the above to be zero.
Z Z
n m 1
z (1 − z) dz = dz
|z|=2 |z|=2 (1 − z)|m| z |n|
In this case we divide the closed curve |z| = 2 into addition of two paths Γ1 and Γ2 where when
ℜ(z) = 12 Γ1 goes from lower half plane to upper half plane along the line ℜ(z) = 12 and Γ2 does the
opposite after that both paths follows the circle perimeter like in the figure
1
ℜ(z) = 2
Γ1
>
>
0 1
Γ2
10
Now (1 − z)−|m| is analytic on Γ1 . Hence
Hence Z Z
n m 1
z (1 − z) dz = dz = 0
|z|=2 |z|=2 (1 − z)|m| z |n|
|dz|
Z Z Z
dz dz
= −iρ = −iρ
|z|=ρ |z − a|4 |z|=ρ z|z − a| 4 2
|z|=ρ z(z − a) (z − a)
2
Z
zdz
= −iρ
|z|=ρ (z − a)2 (zz − az)2
Z
zdz
= −iρ 2 2
|z|=ρ (z − a)2 ρz z − az
Z
zdz
= −iρ
|z|=ρ (z − a)2 (ρ2 − az)2
1
Now if ρ < |a|, then |z − a|4 > 0. Hence the function (z−a)2 (ρ2 −az)2
is analytic and its integral
along |z| = ρ is 0
If ρ > |a| then if ρ2 ̸= az because if it is then
ρ2 ρ2
ρ2 ̸= az ⇐⇒ |z| = ⇐⇒ ρ = ⇐⇒ |a| = ρ
|a| |a|
z
which is not possible. Hence f (z) = (ρ2 −az)2
is analytic in the ρ-disk. Then by Cauchy’s Integral
Formula we have
Z
zdz d z 2πi
2 = 2 2
|z|=ρ (z − a)2 (ρ2 − az) dz (ρ − az) z=a 1!
1 2āa
= + 2πi
(ρ2 − āa)2 (ρ2 − āa)3
ρ2 − |a|2 + 2|a|2 ρ2 + |a|2
= 2πi = 2πi
(ρ2 − |a|2 )3 (ρ2 − |a|2 )3
Hence
ρ2 + |a|2 ρ2 + |a|2
|dz|
Z Z
zdz
= −iρ = −iρ 2πi = 2πρ
|z|=ρ |z − a|4 |z|=ρ (z − a)2 (ρ2 − az)2 (ρ2 − |a|2 )3 (ρ2 − |a|2 )3
11
Problem 12 Ahlfors Page 123: Problem 2
Prove that a function which is analytic in the whole plane and satisfies an inequality |f (z)| < |z|n
for some n and all sufficiently large |z| reduces to a polynomial.
Solution: Given |f (z)| < |z|n . To show f is a polynomial it is enough to show ∃ n ∈ N such that
f (n) (z) = 0. We have Z
n! f (s)ds
f (n) (z) =
2πi |s|=ρ (s − z)n+1
Now if |z| > ρ then
Z
(n+1) (n + 1)! f (t)
f (z) ≤ dt
2π |t|=ρ (t − z)n+2
|f (t)|
Z
(n + 1)!
≤ |dt|
2π |t|=ρ |t − z|n+2
|t|n
Z
(n + 1)!
≤ |dt|
2π |t|=ρ (|t| − |z|)n+2
(n + 1)! ρn ρn+1 (n + 1)!
≤ 2πρ =
2π (ρ − |z|)n+2 (ρ − |z|)n+2
ρn+1 (n+1)!
Hence as |z| → ∞ and ρ → ∞ we have (ρ−|z|)n+2
→ 0. Hence
Solution: We have Z
(n) n! f (s)ds
f (z) =
2πi |s|=ρ (s − z)n+1
Let
Z
(n) n! f (t)
f (z) ≤ dt
2π |t|=R (t − z)n+1
|f (t)|
Z
n!
≤ |dt|
2π |t|=R |t − z|n+1
Z
n! M
≤ |dt|
2π |t|=R (|t| − |z|)n+1
n! M
≤ 2πR
2π (R − |z|)n+1
M n!R M n!R
= ≤
(R − |z|)n+1 (R − ρ)n+1
12
Problem 14 Ahlfors Page 123: Problem 4
If f (z) is analytic for |z| < 1 and |f (z)| ≦ 1/(1−|z|), find the best estimate of f (n) (0) that Cauchy’s
inequality will yield.
Solution: We have Z
(n) n! f (s)ds
f (z) =
2πi |s|=ρ (s − z)n+1
Claim: f (n) (z)
≤ (n + 1)!e
1
Proof: Let for any k ∈ N, |z| = 1 − k = rk . Then
1 1
|f (z)| ≤ = 1 =k
1 − |z| 1− k
Then we have
|f (z)| n!k n+2
Z Z
n! f (z)dz n! n! k kn!
f (n) (0) = ≤ |dz| ≤ n+1 2πrk = n+1 =
2π z n+1 2π |z|n+1 2π rk 1 − k1 (k − 1)n+1
|z|=rk |z|=rk
Hence we have the best estimate of f (n) (0) which is f (n) (0) ≤ (n + 1)!e
Solution: We have Z
(n) n! f (s)ds
f (z) =
2πi |s|=ρ (s − z)n+1
Hence
|f (s)|
Z Z Z
n! f (s)ds n! f (s)ds n!
f (n) (z) = ≤ = |ds|
2πi |s|=ρ (s − z)n+1 2π |s|=ρ (s − z)n+1 2π |s|=ρ |s − z|n+1
13