Chapter 4 Integrals

Complex integral is extremely important, mathematically elegant.

30. Complex-Valued Functions w(t)

First consider derivatives and definite integrals of complex-valued
functions w of a real variable t.

Let w(t )  u(t )  i (t) (1)

Real function of t

w' (t )  dw(t )  u' (t )  i ' (t ) (2)

dt
Provided they exist.

tch-prob 1
 For each complex constant z0  x0  iy0
d  z w(t )   d  ( x  iy ) (u  i ) 
dt  0  dt  0 0 

 d ( x0u  y0 )  i( y0  x0 ) 

dt
 d ( x0u  y0 )  i d ( y0u  x0 )
dt dt
 ( x0  iy0 )(u ' i ')
i.e. d  z0w(t )   z0w'(t )
dt
Various other rules for real-valued functions of t apply here.

e.g. d e z0t  Z e z0t

dt 0

However, not every rule carries over.

tch-prob 2
Example: Suppose w(t) is continuous on a  t  b.
i.e. or (u(t ) and  (t ) are continuous)

The “mean value theorem” for derivatives no longer applies.

There is a number c in a<t<b such that

w' (c)  w(b)  w(a)

ba
a c b
consider w(t )  eit on 0  t  2
w' (t )  ieit  1 never Zero
but w(2 )  w(0)  0

tch-prob 3
Definite Integral of w(t) over a t b

is defined as ab w(t )dt  ab u(t )dt  i ab (t )dt (5)

when exists
Re ab w(t )dt  ab Re  w(t )  dt Imab w(t)dt  ab Im  w(t) dt

Can verify that

b cw(t ) dt  c b w(t )dt

a a
b b b
a (w1(t )  w2 (t )) dt  a w1(t )dt  a w2 (t )dt
b c b
a w(t ) dt  a w(t ) dt  c w(t )dt

tch-prob 4
Anti derivative (Fundamental theorem of calculus)

W (t ) U (t )  iV (t )
if W'(t)  w(t)
then U'(t)  u(t), V'(t)  υ(t)
b b b
a w(t )dt  a u(t )dt  i a (t )dt
 U (t )|ba iV (t )|ba
 U (b)  iV (b)   U (a)  iV (a) 
W (b) W (a)

tch-prob 5
 To Prove ab w(t) dt  ab w(t) dt (a  b)
iθ
Let ab w dt  r0 e 0 b
a w dt  r0
iθ
r0  ab e 0 w dt
        

must be real
real
-iθ
 Re ab e 0 w dt Real part of real number is itself
iθ
 ab Re(e 0 w) dt
iθ iθ iθ
But Re (e 0 w)  e 0 w  e 0 w  w

 r0  ab w dt
or ab wdt  ab w dt

tch-prob 6
31. Contours

Integrals of complex-valued functions of a complex variable are

defined on curves in the complex plane, rather than on just intervals of
the real line.
A set of points z=(x, y) in the complex plane is said to be an arc if

x  x(t) , y  y(t) (a  t  b)

where x(t) and y(t) are continuous functions of real t.

非任意的組合

This definition establishes a continuous mapping of interval a  t  b

into the xy, or z, plane; and the image points are ordered according to in
creasing values of t.

tch-prob 7
It is convenient to describe the points of arc C by

z  z(t ) (a  t  b)

where z(t )  x(t )  i y(t )

The arc C is a simple arc, or a Jordan arc, if it does not cross itself.

i.e. z(t1 )  z(t2 ) when t1  t2

When the arc C is simple except that z(b)=z(a),

we say that C is a simple closed curve, or a Jordan Curve.

tch-prob 8
Ex1:
y
x  ix
 0  x 1
z 

x i


1 x  2 1

0
x
Ex2 : 1 2

z  eiθ ( 0  θ  2π)
Simple closed curve
also z  z0  R eiθ ( 0  θ  2π)

Ex3:
z  eiθ ( 0  θ  2π)

Ex4 :
z  ei2 ( 0  θ  2π)

tch-prob 9
 arc C : z(t)  x(t) i y (t)

Suppose that x’(t) and y’(t) exist and are continuous throughout a  t  b

C is called a differentiable arc

Since z'(t)  x '(t )  iy '(t)

real  valued 2 2
z '(t )  

x '(t )    y '(t )  is integrable over a  t  b
function

The length of the arc is defined as

b
L  z'(t) dt
b
a   z'(t) Δt
a
Δt 0

tch-prob 10
L is invariant under certain changes in the parametric
representation for C

To be specific, Suppose that

t  (τ) (α  τ  β)
where  is a real valued function mapping the interval
(α  τ  β) onto the a  t  b We assume that
interval .
 is continuous with a continuous derivative

We also assume that '(τ)  0 for each τ (t increases with  )

tch-prob 11
 L  αβ z' (τ) '(τ) dτ
Exercise
So., if arc C is
6(b)
z  Z(τ)  z (τ) (α  τ  β)
the fact that
Z'(τ)  z' (τ)'(τ) Exercise 10
allows us to write
L  αβ Z'(τ) dτ

If z  z(t) (a  t  b) is a differentiable arc

and z'(t)  0 in a  t  b
( z'(t) 0　不代表水平，而是在此處
停頓 長度不增加）
Then the unit tangent vector

T  z'(t) is well defined for all t in that open interval.

z'(t)

Such an arc is said to be

smooth.
tch-prob 12
For a smooth arc z  z(t) (a  t  b)

z'(t) is continuous on a  t  b (we agree)

and non zero on a  t  b

A contour, or piecewise smooth arc, is an arc consisting of a

finite number of smooth arcs joined end to end.

If z=z(t) is a contour, z(t) is continuous ,

Whereas z’(t) is piecewise
continuous.
When only initial and final values of z(t) are the same, a
contour is called a simple closed contour

tch-prob 13
32. Contour Integrals
Integrals of complex valued functions f of the complex variable z:
Such an integral is defined in terms of the values f(z) along a given
contour C, extending from a point z=z1 to a point z=z2 in the
complex plane. (a line integral)
Its value depends on contour C as well as the functions f.

z2
Written as c f ( z) dz or z1 f ( z) dz

When value of integral is independent

of the choice of the contour.

Choose to define it in terms  w(t) dt, not as lim Σ .

of

tch-prob 14
Suppose that z  z(t) (a  t  b)

represents a contour C, extending from z1=z(a) to z2=z(b).

Let f(z) be piecewise continuous on C.
Or f [z(t)] is piecewise continuous on a t b

The contour integral of f along C is defined a

f ( z )dz  b f  z(t )  z' (t ) dt (2)
c a  
t 的變化　 define contour C

Since C is a contour, z’(t) is piecewise continuous on a  t  b

Section 31

So the existence of integral (2) is ensured.

tch-prob 15
From section 30
z0: complex constant
c z0 f ( z)dz  z0 c f ( z) dz
and c  f ( z)  g ( z)  dz  c f ( z) dz  c g ( z) dz

Associated with contour C is the contour –C

From z2 to z1

Parametric representation of -C z2=z(b)

z  z(t ) (-b  t  -a)

z1=z(a)

tch-prob 16
 order of C follows a  t  b (t
increasing)
order of –C must also follow increasing parameter value

 b  t  a z(t )  w(t )
小 大 a
b f  w(t )  w' (t ) dt
-t b a
w' (t )  dw(t )  dz(t )  dz(t ) d (t )
dt dt d (t ) dt
  z' (t )
Thus
a
c f ( z) dz  b f  z(t )  

 z' (t ) dt

where z’(-t) denotes the derivative of z(t) with respect

to t, evaluated at –t. or dz(t )  dz(t ) |t t
d (t ) dt

tch-prob 17
After a change of
variable,
τ  t
d  dt
a  τ b
a b
c f(z) dz  b f  z(τ )   z'(τ )  dτ  a f  z(τ )  z'(τ ) dτ
 c f(z) dz

if C  C1  C2 z2
z1 z0
z1  z2 z1  z0 z0  z2
C1 C2
c f(z) dz  c1 f(z) dz  c2 f(z) dz C

Definite integrals in calculus can be interpreted as areas, and they

have other interpretations as well.
Except in special cases, no corresponding helpful interpretation,
geometric or physical, is available for integrals in the complex
plane.

tch-prob 18
33. Examples
 
Ex. 1 I   zdz when C is z  2e i
(    )
C 2 2
 y
By def.
I  2 i i
2e (2e ) ' d
 2
2 C
  0
x
 2 2e i (2iei )d  4i 2 d  4 i -2
2 2

Note for z on the circle z  2, zz  4, or z  4

z
: dz
Hence C z   i.
 

 2 C  2
i i
I  zdz  2
2e (2e ) ' d  
  

  2 2ei (2iei )d  4i 2 ei 2 d  0.
  2  2 
tch-prob 19
 y
Ex2. C1 : contour OAB, C2 : segment OB
A B
1 1+i
C1
f ( z )  y  x  i3 x 2
( z  x  iy ). C1 C2
x
OA : 0
1 i1
z  0  iy
 OA
f ( z )dz   yidy  i  ydy  .
0 0 2 (0  y  1)
1 1 1 1
 f ( z )dz   (1  x  i3x )1dx   (1  x)dx  3i  ( x 2 )dx 
2
- i.
AB 0 0 0 2
AB :
1 i
C1
f ( z )dz 
2
. z  xi
(0  x  1)

1 1
OB :
C2
f ( z )dz   i3 x 2 (1  i )dx  3(1  i )  x 2 dx  1- i.
0 0
z  x  ix
(0  x  1)
1  i
 C1 C 2
f ( z )dz   f ( z )dz  
C1 C2
f ( z )dz 
2
.

tch-prob 20
Ex3 Let C denote an arbitrary smooth arc z=z(t), at b

Want to evaluate I   z dz   z(t) z'(t) dt

C
b

2
d z(t)
Note that  
  z(t) z' (t)
dt 2

z(b)    z(a) 

2 2 2
z(t )  b
 
I 2
 
a 2

 z 2 z
2 2
2 1

z2  z1
2 2
dep. on end points only
 z2
z1 zdz  2
. indep. of the arc.

 c z dz  0 Integral of z around a closed contour in

the plane is zero

tch-prob 21
Ex4. Semicircular path C : z  3eiθ ( 0  θ  π)
C
起點 終點
-3 0 3
Although the branch (sec. 26)
p.77.
f(z)  r eiθ /2 ( r  0, 0  θ  2π)
of the multiple-valued function z1/2 is not defined at the initial
point z=3 of the contour C, the integral
1
I  c z 2 dz of that branch nevertheless exists.

For the integrand is piecewise continuous on C.

z(θ)  3eiθ
f  z(θ)  3 e iθ / 2  3 cos θ  i 3 sin θ ( 0  θ  π)
2 2
right hand limit at θ  0 is 3  i 0.
Hence f  z( )  is continuous on 0  θ  π when its value at θ  0
is defined as 3 . Consequently, I  0 3eiθ / 2 3ie iθ dθ  2 3 (1 i).
tch-prob 22
34. Antiderivatives
－ There are certain functions whose integrals from z1 to z2
　 are independent of path.
　 The theorem below is useful in determining when 　　　
　　 integration is independent of path and, moreover, when a
n 　 integral around a closed path has value zero.
－ Antiderivative of a continuous function f : a function F such
that F’(z)=f(z) for all z in a domain D.
－ note that F is an analytic function.

tch-prob 23
Theorem: Suppose f is continuous on a domain D.
The following three statements are equivalent.
(a) f has an antiderivative F in D.
(b) The integrals of f(z) along contours lying entirely in D
and extending from any fixed point z1 to any fixed point
z2 all have the same value.
(c) The integrals of f(z) around closed contours lying
entirely in D all have value zero.
Note: The theorem does not claim that any of these statements
is true for a given f in a given domain D.

tch-prob 24
Pf
: A:(a)  (b)
(a) f has am antiderivative F.
d F  z(t)  F'  z(t) z'(t)  f  z(t) z'(t)
dt      

b b
c f(z) dz  a f  z(t) z'(t)dt  F  z(t) a  F  z(b)  F  z(a)
 F(z2 )  F(z1 )
B:(b)  (c)
C  C1  C2 (b
c f(z) dz  c f(z)dz )
1 2
z2
c f(z)dz  c f(z)dz  0 c f(z)dz  c f(z)dz  0 C2
1 2 1 2
c f(z)dz  0 C1
z1
(c)
C :(c)  (b)
c f(z) dz  0 c1  -c2  0 c1 f(z) dz  c2 f(z) dz
(b
)
tch-prob 25
(c) -->(a)
Define F ( z)  zz f (s) ds
0

F ( z z)  F ( z)  zzz f (s)ds  zz f (s) ds  zzz f (s)ds

0 0
zz ds  z (choose integration path from z to z  z as
z
a line segment)
f(z)  1 zzz ds f ( z)  1 zzz f ( z)ds
z z
F ( z z) - F ( z)  f ( z)  1 zz  f (s) - f ( z)  ds
z z z  
Since f is continuous,
f(s)- f(z)  ε whenever s- z  δ
F ( z z)-F ( z)
 f ( z)  1 zz z f (s)  f ( z) ds  1 ε z  ε
z z z
when z z is close to z, so that z  δ
lim
 z 0 F ( z z)  f ( z) or F'(z)  f(z)
z

tch-prob 26
35. Examples
Ex1. f ( z )  z 2 continuous
has an antiderivative F ( z )  1 z 3 throughout the plane.
3
3 1 i
1+i z 1 2
Hence 0
z 2 dz 
3
 (1  i )3  (1  i )
3 3
0

for every contour from z  0 to z  1  i.

tch-prob 27
Ex2. 12 is continuous every where except at the origin.
z
It has an antiderivative - 1 in the domain z  0.
z
dz 1 z2 1 1
consequently, z
z12    (z1  0 , z2  0 )
z 2 z
z 1 z1 z2

For any contour from z1 to z2 that does not pass through the
origin . dz  0 for C: z  2ei (-    )
c z 2

Note that: f ( z)  1z can not be evaluated in a similar way

1
though derivative of any branch F(z) of log z is z ,
F(z) is not differentiable, or even defined, along its branch cut.
(p.77)
(In particular , if a ray    from the origin is used to form
the branch cut, F'(z) fails to exist at the point wheree the ray
intersect the circle C.  C does not lie in a domain throughout
1
which F '( z )  .
z tch-prob 28
 excluded
Ex3.
Let D : z  0,    Argz   .
The principal branch Log z of the logarithmic function serves as
an antiderivative of the continuous function 1/z throughout D.
Hence
2i dz 2i
2i  Log z  Log(2i)  Log(2i)
z 2 i
 
 (ln 2  i )  (ln 2  i )   i (- π  θ  π )
2 2 2 2
when the path is the arc z  2eiθ (- π  θ  π )
2 2

(compare with p.98)

tch-prob 29
Ex4.
1
Evaluate  z C1
2 dz
C1
1
where z is the branch
2

1 iθ
z 2 r e 2 -3 3
(r  0, 0  θ  2π) C2
(2)

C1 is any contour from z=-3 to z=3, that lies above the

x-axis. (Except end points)

tch-prob 30
1. The integrand is piecewise continuous on C1, an
d the integral therefore exists.
2. The branch (2) of z 1/2 is not defined on the ray
in particular at the point z=3. F(z) 不可積
3. But another branch.
iθ
F1(z)  r e 2 (r  0, - π  θ  3π )
2 2
is defined and continuous everywhere on C1.
4. The values of F1(z) at all points on C1 except z=3
coincide with those of our integrand (2); so the in
tegrand can be replaced by F1(z).

tch-prob 31
Since an antiderivative of F1(z) is
i3θ
F (z)  z 2  r r e 2 (r  0, - π  θ  3π )
2 2
3

1 3 3 2 2

We can write
1 3 3 i3π
i 0
dz  -3 f1(z) dz  F1(z)  2 3 (e  e 2 )
c1 z 2
3
 2 3 (1 i)

(cf. p. 100, Ex4)

1
For c z 2 dz
2 -3 3

Replace the integrand by the branch

tch-prob 32
 iθ
f 2(z)  r e 2 (r  0, π  θ  5π )
2 2
i3θ
Then F2(z)  r r e 2 (r  0, π  θ  5π )
2
3 2 2
is an antiderivative of f 2(z).
1 3 3 i3π
i3π
c2 z 2dz  3 f 2(z) dz  F2(z) 3  2 3 (e  e 2 )

 2 3 (-1 i)
1
c2 c1 z 2 dz  -4 3
0

c  c  two contours can not be on the some domain

1 2

tch-prob 33
36. Cauchy-Goursat Theorem

We present a theorem giving other conditions on a function

f ensuring that the value of the integral of f(z) around a
simple closed contour is zero.
Let C denote a simple closed contour z=z(t) (a  t  b),
described in the positive sense (counter clockwise).
Assume f is analytic at each point interior to and on C.

b
c f ( z) dz  a f  z(t )  z' (t ) dt
if f ( z)  u( x,y)  iv( x,y), and z  x(t )  i y(t ).
b b
c f ( z) dz  a (ux'  vy' )dt  i a (v x'  uy' )dt.

Green's theorem: If P( x, y),Q( x, y), Px , Py ,Qx ,Qy are continuous throughout

the closed region R consisting of all points interior to and on C,
c Pdx  Qdy  R (Qx  Py ) dA
tch-prob 34
 then C f ( z)dz  R (vx u y )dA  i R (u x v y )dA
But
ux  v y u y  vx (Cauchy-Riemann eg.)

when f is analytic in R and f' is continuous there

c f ( z) dz  0  Cauchy 's finding.

Goursat was the first to prove that the condition of continuity o

n f’ can be omitted.
Cauchy-Goursat Theorem:
If f is analytic at all points interior to and on a simple c
losed contour C, then

c f ( z) dz  0

tch-prob 35
37. Proof: Omit
38. Simply and Multiply Connected Domains
A simply connected domain D is a domain such that every simple
closed contour within it encloses only points of D.
Multiply connected domain : not simply connected.
Can extend Cauchy-Goursat theorem to:
Thm 1: If a function f is analytic throughout a simply connected d
omain D, then
c f ( z) dz  0

for every closed contour C lying in D. 分成n

simple
Not just simple closed contour as before. closed
contours

tch-prob 36
Corollary 1. A function f which is analytic throughout a simply connect
ed domain D must have an antiderivative in D.
pf : section 34 for continuous f
given
f analytic c f ( z) dz  0
 f continuous  f has antiderivative
Extend cauchy-goursat theorem to boundary of multiply connected domain

Theorem 2. If f is analytic within C and on C except for points

interior to any Ck, ( which is interior to C, ) then
n C
c f ( z ) dz   c f ( z)dz  0 C1 C2
k 1 k
C: simple closed contour, counter clockwise
Ck: Simple closed contour, clockwise

tch-prob 37
Corollary 2. C1

Let C1 and C2 denote positively oriented simple

C2
closed contours, where C2 is interior to C1.
If f is analytic in the closed region consisting of
those contours and all points between them, then

c1 f ( z) dz  c2 f ( z) dz Principle of

pf : from thm 2, c f ( z) dz  -c f ( z)dz  0
1 2 deformation of paths.

tch-prob 38
Example:
C

C is any positively oriented simple closed contour

surrounding the origin. C0

dz
Want to show that   2 i.
C z

Construct a positively oriented circle C0 with center at the

origin and radius so small that C0 lies entirely inside C.
i
dz 2 i  e
Since   i
d  2 i
C0 z 0 e
and since 1/ z is analytic everywhere except at z =0,
dz
we have   2 i from corollary 2.
C z

2 2
  zdz    e i  e d  i  
i i
zdz  2
ei 2 d  0
C C0 0 0

C
z 2 dz  ?
tch-prob 39
39. Cauchy Integral Formula
Thm. Let f be analytic everywhere within and on a simple close
d contour C, taken in the positive sense. If z0 is any point interio
r to C then,
f ( z0 )  1 c f z(-z)zdz (1) Cauchy integral formula
2π i 0
or (Values of f interior to C are
f ( z)dz  2π i f ( z ) completely determined by the
c z  z 0 values of f on C)
0

Ex. Let C be z  2
since f ( z)  z 2
is analytic within and on C,
9 z
and since -i is interior to C.
z
z dz 9 z 2 -i π
c (9 - z 2 )( z  i)  c z  (i) dz  2π i (10)  5

tch-prob 40
Pf. of theorem: C

C0
f is continuous at z0
f ( z)  f ( z0 )   when z - z0   z0

choose a positive   
C0 : z - z0  

since
f ( z)
z  z0 is analytic in the closed region consisting of C
and C0 and all points between them, from corollary 2,
section 38,
f ( z)dz f ( z) dz
c z  z  c0 z  z
0 0

c f z(z) dz  f ( z ) dz  f ( z)  f ( z0 ) dz
z0 0 c z  z c0 z  z0
0 0

tch-prob 41
but c z dz
 z  2 i
0 0
f ( z) - f ( z ) f ( z)  f ( z0 )
 c z  z 0 - 2π i f ( z0 )  c z  z dz
0 0 0
f ( z)  f ( z0 ) f ( z)  f ( z0 )
also c dz  c dz
0 z  z0 0 z  z0

f ( z)  f ( z0 )
 c dz   2πρ
0 z  z0 
 2π

 c fz(z)zdz  2π i f ( z0 )  2π
0 0

Non-negative constant arbitrary

c f z(
z) dz  2π i f ( z )  0
z0 0

tch-prob 42
40. Derivatives of Analytic Functions
To prove : f analytic at a point
its derivatives of all orders exist at that point
and are themselves analytic there.
z0  z interior to C
in Cauchy integral formula
zs on C S C

z
f ( z)  1 c f s(s) zds
d
z
2π i

f ( z z)  f ( z)  1 ( 1  1 ) f (s) ds
z 2π i c s  z z s  z z when 0  z  d ,
 1 c f (s) ds where d is the
2π i (s  z z)(s  z)
shortest distame
from z to s on C.
tch-prob 43
  
but c 1  1  f (s) ds

(s  z z)(s  z) (s  z) 
2

 z c f (s) ds
(s  z z)(s  z)2
Let M denote the maximum value of f (s) on C.
Let L be the length of C.
Note that s - z  d
s  z z  s  z  z  d - z
f (s) ds z ML
 z c   0 as z  0
(s  z z)(s  z) (d - z )d
2 2

lim f ( z z) - f ( z)  1 f (s) ds


z  0 z 2π i c (s  z)2 #
 f '( z)  1 c f (s) ds2
2π i (s  z)
similarly for f '' ( z)  1 c f (s) ds3
π i (s  z )


tch-prob 44
Thm1. If f is analytic at a point, then its derivatives of all orders are
also analytic functions at that pint.
In particular, when
f ( z)  u( x, y)  iv( x, y) is analytic at a point z  ( x,y)

the analyticity of f ' eusures the continuity of f ' there

f '( z)  ux ( x,y)  i vx ( x,y)  v y ( x,y)  i u y ( x,y)
 ux , u y , vx , v y are continuous at (x,y)

similarly
f ''( z)  uxx(x,y)  i vxx ( x,y)  v yx ( x,y)  i u yx ( x,y)

Corollary : If f ( z)  u( x,y)  i v( x,y) is analytic at a point

z  x  iy, then u, v have continuous partial
derivatives of all orders at that point.

tch-prob 45
Let f (0) ( z) denote f ( z), and 0!1, can use mathematical
induction to verify that
f (n) ( z)  n! c f (s)(ds (n  0,1,2....)
2π i (s  z) n 1)

or c f ( z) (dz  2π i f (n) ( z )
( z  z0 ) n1) n! 0

Ex1: C : z 1 f ( z )  e2 z
e2 z dz  f ( z) dz  2π i f''' (0)  8π i
c z 4 c ( z  0)31 3! 3

Ex2: z0 in C, f ( z) 1,
dz  2π i dz  0 (n 1, 2, ....)
c z- z c ( z - z )n 1
0 0

Thm 2: If f continuous in domain D and

c f ( z) dz  0 for every closed contour C in D,
then f is analytic throughout D.

tch-prob 46
41. Liouville’s Theorem and the Fundamental Theorem of
Algebra
Let z0 be a fixed complex number.
If f is analytic within and on a circle z  z0  R

then f (n) ( z0 )  n! c f ( z) dz (n 1, 2,.......).

2πi ( z  z0 )n1
Let MR denote the Maximum value of f (z)

f (n) ( z0 )  n! MnR1 2πR (n 1, 2,....)

2π R
n! M
f (n) ( z0 )  n R , (n 1, 2,.....) Cauchy's inequality.
R
M
for n 1 f '( z0 )  R
R

tch-prob 47
Thm 1 (Liouville’s theorem): If f is entire and bounded in the comp
lex plane, then f(z) is constant throughout the plane.
finite
pf : f '( z0 )  M
R
可以 Arbitrarily large
for R large, f '( z0 ) must be zero
since z0 is arbitrary,
f '( z)  0 everywhere
f is a constant
Thm 2 (Fundamental theorem of algebra):
Any polynomial
P( z)  a0  a1z  a2 z 2  ......... an z n (an  0)
of degree n (n 1) has at least one zero.
i.e., there exists at least one point z0 s.t. P( z0 )  0
Pf. by contradiction
tch-prob 48
Suppose that P(z) is not zero for any value of z.
f ( z )  1
Then P( z) is clearly entire and it is
also bounded in the complex plane.
To show that it is bounded, first write
a0 an 1
a1 a2
w .... z
n 1

z z z n2n

So that P( z)  (an  w) z n
Can find a sufficiently large positive R such that
ai an
 , i  0,1, ........n-1, when z  R
z n i 2n
a0 a1 a2 an1 a0 a1 an1
w    ....    .....
z n z n 1 zn2 z z n z n 1 z

Generalized triangle inequality

an
 when z  R
2
tch-prob 49
 an
an  w  an  w 
2
a a
P( z)  an  w z n  n z n  n Rn z R
2 2
Then f ( z)  1  2 n z R
P( z) an R

So, f is bounded in the region exterior to z  R.

But f is continuous in z  R  f is bounded in z  R.
 f is bounded in the entire plane.
From Liouville's theorem, f ( z ) is constant.
 P( z ) is constant.
But P( z ) is not constant  We have a contradiction.

tch-prob 50
From the (F. T. 0. A) theorem
any polynomial P(z) of degree n can be expressed as
P( z)  ( z  z1) Q1( z) (Ex.10, sec 42 )

Polynomial of degree n-1

but Q1( z)  ( z  z2 ) Q2 ( z)

Polynomial of degree n-2

P( z)  ( z  z1)( z  z2 ) Q2 ( z)
 :
:
 c ( z  z1)( z  z2 ).......( z  zn )

tch-prob 51
42. Maximum Moduli of Functions
Lemma. Suppose that f(z) is analytic throughout a neighborhood
z  z0  ε of a point z0. If f ( z)  f ( z0 )
for each point z in that neighborhood, then f(z) has the
constant value f(z0) throughout the neighborhood .
Pf : Let z1 be any point other than z0 in the neighborhood.
Let  be the distance between z1 and z0 .
C
C : circle z  z0   . z1
z0

f ( z0 )  1 c f z(
z) dz
z0
2π i  
iθ
f ( z0  ρe )
 1 02 i ρ eiθ dθ
2π i ρe iθ

1 Cρ : z  z0  ρeiθ ( 0  θ  2π)
 0 f ( z0  ρe ) dθ
2 iθ
2π
f’s value at the center is the arithmetic mean of its values on the
circle. ~ Gauss’s mean value theorem.
tch-prob 52
 f ( z0 )  1 02 f ( z0  ρei ) d ..........(3)
2π
On the other hand, since f ( z0  ρei )  f ( z0 ) ..........(4)
i
0 f ( z0  ρe ) d  0 f ( z0 ) dθ  2π f ( z0 )
2 2

 f ( z0 )  1 02 f ( z0  ρei ) d ......(5)

2π

From (3) and (5)

f ( z0 )  1 02 f ( z0  ρei ) dθ
2π
or 02  f ( z0 )  f ( z0  ρei )  dθ  0
        
 0 from (4), and continuous in 

tch-prob 53
 f ( z0 )  f ( z0  ρei ) ( 0  θ  2π)
 f ( z)  f ( z0 ) for all points z in z - z0  ρ
 f ( z)  f ( z0 ) for all points z in z - z0  ε
From Ex 7(c), sec 22, (when the modulus of an analytic function
is constant in a domain, the function itself is constant there.)
 f ( z)  f ( z0 ) for all points z in the neighborhood.

Thm. (maximum modulus principle)

If a function f is analytic and not constant in a given domain
D, then f (z) has no maximum value in D. That is, there i
s no point z0 in the domain such that f ( z)  f ( z0 ) for all
points z in it.

tch-prob 54
 Pf:
D

N2 N n-1
N1
N0
z2 Nn
z0 z1 zn-1 zn

p
d

D is a connected open set.

 There is a polygonal line L lying entirely in D that extends from z0 , to any
other point P in D.
Let d be the shortest distance from points on L to the boundary of D.
We then form a finite sequence of points z0 , z1, z2 ,  , zn (=P ) along L, where each point
is sufficiently close to the adjacent ones that
zk  zk 1  d (k  1, 2, , n)

Construct Nk : z  zk  d
zk is in zk -1's neighborhood
tch-prob 55
Assume f(z) has a max value in D at z0.
f(z) also has a max value in N0 at z0.

From Lemma, f(z) has constant value f(z0) throughout N0.

 f ( z1)  f ( z0 ) ( z1 in N0 )

 f ( z)  f ( z1) for all z in N1

from Lemma  f ( z)  f ( z1)  f ( z0 )

 f ( z2 )  f ( z0 ) (z2 in N1)
:
 f ( zn )  f ( z0 ) for every z in D
 f ( z) is constant  a contradiction
 f ( z) has no maximum value in D

tch-prob 56
 If a function f that is analytic at each point in the interior
of a closed bounded region R is also continuous
throughout R, then the modulus f (z) has a maximum
value somewhere in R. (sec 14) p.41
i.e..
f ( z)  M for all points z in R.

If f is a constant fumction,
then f ( z)  M for all z in R.
Maximum at the boundary. not continuous
on boundary

Corollary: Suppose f is continuous in a closed bounded region R.

and that it is analytic and not constant in the interior of R.

Then Maximum value of f (z) in R occurs somewhere on the

boundary R and never in the interior.

tch-prob 57