Complex Analysis Fall 2007 Homework 4: Solutions

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Complex Analysis

Fall 2007 Homework 4: Solutions

1.5.2.

(a) The function f (z) = 3z 2 +7z+5 is a polynomial so is analytic everywhere with derivative
f 0 (z) = 6z + 7.
(b) The function f (z) = (2z + 3)4 is a composition of polynomials so is analytic everywhere
with derivative f 0 (z) = 8(2z + 3)3 (by the chain rule).
(c) The function f (z) = (3z − 1)/(3 − z) is rational and so is analytic where z 6= 3. By the
quotient rule, its derivative is
8
f 0 (z) = .
(3 − z)2
p
1.5.10. Let f (z) = |z| and write f = u + iv and z = x + iy. Then u = x2 + y 2 and v = 0.
We find that
∂u x ∂u y
=p , =p
∂x x2 + y 2 ∂y x2 + y 2
∂v ∂v
= = 0.
∂x ∂y
In order for f to be analytic the Cauchy-Riemann equations must hold. That is, we need
x y
p =p =0
x2 + y2 x2 + y2
which is impossible, since it requires x and y to be both simultaneously zero and nonzero.
Therefore f is differentiable nowhere.
p
Here’s another proof. Since the partial derivatives of Re(|z|) = x2 + y 2 do not exist
at (0, 0), the Cauchy-Riemann theorem implies that |z| cannot be differentiable at z = 0.
Suppose that f (z) = |z| were analytic at some z0 ∈ C, z0 6= 0. Then (f (z))2 = |z|2 = zz
would be analytic at z0 and, since z0 is nonzero, so too would be (f (z))2 /z = z. But we
know that g(z) = z is analytic nowhere, so this is impossible. Therefore f (z) = |z| cannot
be analytic at any nonzero complex number either. Hence, |z| is differentiable nowhere, i.e.
is not analytic.
1.5.14.

(a) According to Cauchy-Riemann equations, if f is analytic then


∂f 1 ∂f
f0 = =
∂x i ∂y

1
so that  
∂f 1 ∂f 1 ∂f 1
= + = (f 0 + f 0 ) = f 0 .
∂z 2 ∂x i ∂y 2
(b) If f (z) = z = x + iy then f is analytic so that by part (a) we have

∂f z = f 0 = 1.
Moreover
∂f ∂f
= 1, =i
∂x ∂y
so that  
∂f 1 ∂f 1f 1
= − = (1 − 1) = 0.
∂z 2 ∂x iy 2
(c) If f (z) = z = x − iy then
∂f ∂f
= 1, = −i.
∂x ∂y
Therefore  
∂f 1 ∂f 1 ∂f 1
= + = (1 − 1) = 0
∂z 2 ∂x i ∂y 2
and  
∂f 1 ∂f 1 ∂f 1
= − = (1 + 1) = 1.
∂z 2 ∂x i ∂y 2
∂ ∂
(d) We show first that ∂x and ∂y satisfy the sum, product and scalar multiple rules when
applied to complex valued functions. Let f, g be complex valued functions and let c be
a complex scalar. Write f = s + it, g = u + iv, c = a + ib.
Scalar multiplication. We have cf = (as − bt) + i(at + bs) so that by definition and the
linearity of derivative of real functions we have
∂(cf ) ∂(as − bt) ∂(at + bs)
= +i
∂x ∂x  ∂x 
∂s ∂t ∂t ∂s
= a −b +i a +b
∂x ∂x ∂x ∂x
 
∂s ∂t
= (a + ib) +i
∂x ∂x
∂f
= c
∂x
∂ ∂
which proves the scalar multiplication rule for ∂x
. The proof for ∂y
is identical, replacing
x by y throughout.

2
Sum. We have f + g = (s + u) + i(t + v) so that by definition and the linearity of the
derivative of real functions we have
∂(f + g) ∂(s + u) ∂(t + v)
= +i
∂x ∂x  ∂x 
∂s ∂u ∂t ∂v
= + +i +
∂x ∂x ∂x ∂x
   
∂s ∂t ∂u ∂v
= +i + +i
∂x ∂x ∂x ∂x
∂f ∂g
= +
∂x ∂x
∂ ∂
which proves the addition rule for ∂x
. The proof for ∂y
is identical, replacing x with y
throughout.
Product. We have f g = (su − tv) + i(tu + sv) so that by definition and the product rule
for derivatives of real functions we have
∂(f g) ∂(su − tv) ∂(tu + sv)
= +i
∂x ∂x ∂x  
∂u ∂s ∂v ∂t ∂u ∂t ∂v ∂s
= s +u −t −v +i t +u +s +v
∂x ∂x ∂x ∂x ∂x ∂x ∂x ∂x
       
∂u ∂v ∂u ∂v ∂s ∂t ∂t ∂s
= s −t +i t +s + u −v +i u +v
∂x ∂x ∂x ∂x ∂x ∂x ∂x ∂x
   
∂u ∂v ∂s ∂t
= (s + it) +i + (u + iv) +i
∂x ∂x ∂x ∂x
∂g ∂f
= f +g
∂x ∂x
∂ ∂
which proves the product rule for ∂x
. The proof for ∂y
is identical, replacing x with y
throughout.
Finally, we are in a position to verify that the sum, product and scalar multiplication
∂ ∂
rules hold for ∂z and ∂z . We prove a slightly more general statement that includes both
of these statements as special cases.
∂ ∂
Let a, b ∈ C and set D = c ∂x + b ∂y . I claim that D satisfies the sum, product and scalar
multiple rules. Let f, g be complex functions and let c ∈ C. By what we have already
shown we have
∂(cf ) ∂(cf )
D(cf ) = a +b
∂x ∂y
∂f ∂f
= ac + bc
∂x ∂y
= cD(f )

3
which proves the scalar multiple rule. Also
∂(f + g) ∂(f + g)
D(f + g) = a +b
∂x ∂y
   
∂f ∂g ∂f ∂g
= a + +b +
∂x ∂x ∂y ∂y
∂f ∂f ∂g ∂g
= a +b +a +b
∂x ∂y ∂x ∂y
= D(f ) + D(g),

proving the sum rule. Finally


∂(f g) ∂(f g)
D(f g) = a +b
∂x ∂y
   
∂g ∂f ∂g ∂f
= a f +g +b f +g
∂x ∂x ∂y ∂y
   
∂g ∂g ∂f ∂f
= f a +b +g a +b
∂x ∂y ∂x ∂y
= f D(g) + gD(f ),

which establishes the product rule for D.


Now we are finished since
∂ 1 ∂ 1 ∂
= +
∂z 2 ∂x 2i ∂y
and
∂ 1 ∂ 1 ∂
= −
∂z 2 ∂x 2i ∂y
are both of the form dealt with above.
(e) (⇐) This is the easy direction. If anm = 0 whenever m 6= 0 we have
N X
X M N
X N
X
n m n 0
anm z z = an0 z z = an0 z n
n=0 m=0 n=0 n=0

which is a polynomial in z and hence analytic everywhere.


(⇒) For this we use the result of exercise 13, which says that if f (z) is analytic then
∂f
= 0.
∂z
∂ ∂z m
Since ∂z obeys the product rule and ∂z = 1 an easy induction shows that ∂z∂z = mz m−1
n
for m ∈ Z+ . Moreover, since z n is analytic we know that ∂z∂z
= 0. Therefore, if the

4
expression in question is analytic we have (by the sum, product and scalar multiple
rules)
N M
!
∂ XX
0 = anm z n z m
∂z n=0 m=0
N X M
X ∂
= (anm z n z m )
n=0 m=0
∂z
N X M  m n

n ∂z m ∂z
X
= anm z +z
n=0 m=0
∂z ∂z
N X
X M
= anm mz n z m−1 z m .
n=0 m=0

But the only way that a polynomial can be identically zero is if all of its coefficients
vanish. That is, we need manm = 0 for all m and n. In particular, if m 6= 0 we must
have anm = 0, which is exactly what we sought to prove.

1.5.16. Before we begin, notice that the statement au(x, y) + bv(x, y) = c with not all of
a, b, c equal to zero is equivalent to saying that the values of f (z) = u(x, y) + iv(x, y) lie on
a straight line.

(a) We write u = u(x, y), v = v(x, y) etc. to simplify notation. If we apply ∂/∂x and ∂/∂y
to the equation au + bv = c we obtain
∂u ∂v
a +b = 0
∂x ∂x
∂u ∂v
a +b = 0.
∂y ∂y
Since f is analytic we can apply the Cauchy-Riemann equations in the second equality
above to obtain the system
∂u ∂v
a +b = 0
∂x ∂x
∂u ∂v
b −a = 0
∂x ∂x
or, in matrix form,
∂u
 
a b ∂x
0
∂v = .
b −a ∂x
0
The determinant of the coefficient matrix is −(ab + b2 ) which cannot be zero because
a, b, c are real and not all zero. Hence, the only solution to the system is
∂u ∂v
= = 0.
∂x ∂x

5
Therefore
∂u ∂v
f 0 (z) =
+i =0
∂x ∂x
and, since A is connected, we conclude that f is constant.
(b) If a, b, c are complex, then the proof above does not apply. However, if we write a =
a1 + ia2 , b = b1 + ib2 and c = c1 + ic2 with a1 , a2 , b1 , b2 , c1 , c2 ∈ R, then the equation
au + bv = c is equivalent to the pair of equations
a1 u + b 1 v = c 1
a2 u + b 2 v = c 2 .
Since not all of a, b, c are zero, it must be that in at least one of these equations not
all of the constants are zero. We may then apply part (a) to that equation to conclude
that f is constant. In other words, the statement is valid for complex a, b, c as well.

1.5.20. Let A ⊂ C be connected and open and let f : A → C be an analytic function. We


are asked to prove the following: if f (n + 1) = 0 for some n ∈ Z+
0 then f is a polynomial of
degree ≤ n. We induct on n.
When n = 0 the statement we need to prove is: if f 0 = 0 then f is constant. This is
Proposition 1.5.5 and was proven in class. We now assume the statement holds for some
n ≥ 0. Suppose that f (n+2) = 0. Then (f 0 )(n+1) = f (n+2) = 0 so that, by our assumption, f 0
is a polynomial of degree ≤ n. Write
n
X
f 0 (z) = ak z k
k=0

with all ak ∈ C. Let


n
X ak k+1
h(z) = f (z) − z .
k=0
k+1
Then h(z) is analytic on A and
n
X
0 0
h (z) = f (z) − ak z k = 0.
k=0

But we have already seen that this implies h0 is constant. If h0 = c ∈ C then we see that
n n
X ak k+1 X ak k+1
f (z) = h(z) + z =c+ z ,
k=0
k+1 k=0
k+1

i.e. f is a polynomial of degree ≤ n + 1. Therefore, if the statement “f (n + 1) = 0 implies


f is a polynomial of degree ≤ n” holds for any n ≥ 0 it also holds for n + 1. By induction,
we conclude that the statement is true for all n ≥ 0.

You might also like