Let
G
⊆
C
{\displaystyle G\subseteq \mathbb {C} }
be open,
f
:
G
→
C
{\displaystyle f\colon G\to \mathbb {C} }
holomorphic,
z
0
∈
G
{\displaystyle z_{0}\in G}
a point in
G
{\displaystyle G}
, and
U
:=
D
r
(
z
0
)
⊂
G
{\displaystyle U:=D_{r}(z_{0})\subset G}
a bounded disk in
G
{\displaystyle G}
. Then for all
z
∈
D
r
(
z
0
)
{\displaystyle z\in D_{r}(z_{0})}
(i.e., for all
z
{\displaystyle z}
with
|
z
−
z
0
|
<
r
{\displaystyle |z-z_{0}|<r}
), the following holds: :
f
(
z
)
=
1
2
π
i
∮
∂
U
f
(
ζ
)
ζ
−
z
d
ζ
{\displaystyle f(z)={\frac {1}{2\pi \mathrm {i} }}\oint _{\partial U}{\frac {f(\zeta )}{\zeta -z}}\mathrm {d} \zeta }
Here,
∂
U
{\displaystyle \partial U}
denotes the positively oriented curve
t
↦
z
0
+
r
e
i
t
{\displaystyle t\mapsto z_{0}+re^{\mathrm {i} t}}
for
t
∈
[
0
,
2
π
]
{\displaystyle t\in [0,2\pi ]}
along the boundary of the disk
U
{\displaystyle U}
.
For a fixed
z
∈
U
{\displaystyle z\in U}
, the function
g
:
U
→
C
{\displaystyle g\colon U\to \mathbb {C} }
defined by
w
↦
f
(
w
)
−
f
(
z
)
w
−
z
{\displaystyle w\mapsto {\tfrac {f(w)-f(z)}{w-z}}}
for
w
≠
z
{\displaystyle w\neq z}
und
w
↦
f
′
(
z
)
{\displaystyle w\mapsto f'(z)}
for
w
=
z
{\displaystyle w=z}
.
g
{\displaystyle g}
is steadily on
U
{\displaystyle U}
and holomorphic on
U
∖
{
z
}
{\displaystyle U\setminus \{z\}}
. By the Cauchy Integral Theorem , we now have: :
0
=
∮
∂
U
g
(
w
)
d
w
=
∮
∂
U
f
(
ζ
)
ζ
−
z
d
ζ
−
f
(
z
)
∮
∂
U
d
ζ
ζ
−
z
{\displaystyle 0=\oint _{\partial U}g(w)dw=\oint _{\partial U}{\frac {f(\zeta )}{\zeta -z}}\mathrm {d} \zeta -f(z)\oint _{\partial U}{\frac {\mathrm {d} \zeta }{\zeta -z}}}
.
The function
h
:
U
→
C
{\displaystyle h\colon U\to \mathbb {C} }
,
w
↦
∮
∂
U
d
ζ
ζ
−
w
{\displaystyle \textstyle w\mapsto \oint _{\partial U}{\frac {\mathrm {d} \zeta }{\zeta -w}}}
is holomorphic with the derivative
h
′
(
w
)
=
∮
∂
U
d
ζ
(
ζ
−
w
)
2
{\displaystyle \textstyle h'(w)=\oint _{\partial U}{\frac {\mathrm {d} \zeta }{\left(\zeta -w\right)^{2}}}}
, which vanishes since the integrand has an antiderivative (namely
ζ
↦
−
1
ζ
−
w
{\displaystyle \zeta \mapsto -{\frac {1}{\zeta -w}}}
). Therefore,
h
{\displaystyle h}
is constant, and since
h
(
a
)
=
2
π
i
{\displaystyle h(a)=2\pi i}
, we have
h
(
z
)
=
2
π
i
{\displaystyle h(z)=2\pi i}
.
Consequences of the Cauchy Integral Theorem
edit
The Cauchy Integral Theorem (CIS) leads to the following corollaries:
Representation of the Function at the Center of the Disk
edit
For every holomorphic function, the function value at the center of a circle is the average of the function values on the circle's boundary. Use
ζ
(
t
)
=
z
o
+
r
e
i
t
,
d
ζ
=
i
r
e
i
t
d
t
{\displaystyle \zeta (t)=z_{o}+re^{\mathrm {i} t},\ \mathrm {d} \zeta =\mathrm {i} re^{\mathrm {i} t}\mathrm {d} t}
.
Test: :
f
|
U
(
z
o
)
=
1
2
π
i
∮
∂
U
f
(
ζ
)
ζ
−
z
o
d
ζ
=
1
2
π
i
∫
0
2
π
f
(
z
o
+
r
e
i
t
)
r
e
i
t
i
r
e
i
t
,
d
t
=
1
2
π
∫
0
2
π
f
(
z
o
+
r
e
i
t
)
,
d
t
{\displaystyle {\begin{aligned}f|{U}(z_{o})&={\frac {1}{2\pi \mathrm {i} }}\oint {\partial U}{\frac {f(\zeta )}{\zeta -z_{o}}}\mathrm {d} \zeta ={\frac {1}{2\pi \mathrm {i} }}\int _{0}^{2\pi }{\frac {f(z_{o}+re^{\mathrm {i} t})}{re^{\mathrm {i} t}}}\mathrm {i} re^{\mathrm {i} t},\mathrm {d} t\ &={\frac {1}{2\pi }}\int _{0}^{2\pi }f(z_{o}+re^{\mathrm {i} t}),\mathrm {d} t\end{aligned}}}
Every holomorphic function is infinitely complex differentiable, and each of these derivatives is also holomorphic. Expressed using the integral formula, this means for
|
z
−
z
o
|
<
r
{\displaystyle |z-z_{o}|<r}
and
n
∈
N
0
{\displaystyle n\in \mathbb {N} {0}}
: :
f
(
n
)
(
z
)
=
n
!
2
π
i
∮
∂
U
f
(
ζ
)
(
ζ
−
z
)
n
+
1
d
ζ
.
{\displaystyle f^{(n)}(z)={\frac {n!}{2\pi \mathrm {i} }}\oint {\partial U}{\frac {f(\zeta )}{(\zeta -z)^{n+1}}}\mathrm {d} \zeta .}
Local Developability in Power Series
edit
Every holomorphic function can be locally expanded into a power series for
|
z
−
z
o
|
<
r
{\displaystyle |z-z_{o}|<r}
.
f
(
z
)
=
∑
n
=
0
∞
(
1
2
π
i
∮
∂
U
f
(
ζ
)
(
ζ
−
z
o
)
n
+
1
d
ζ
)
(
z
−
z
o
)
n
=
∑
n
=
0
∞
a
n
(
z
−
z
o
)
n
.
{\displaystyle f(z)=\sum \limits _{n=0}^{\infty }\left({\frac {1}{2\pi \mathrm {i} }}\oint _{\partial U}{\frac {f(\zeta )}{(\zeta -z_{o})^{n+1}}}\mathrm {d} \zeta \right)(z-z_{o})^{n}=\sum \limits _{n=0}^{\infty }a_{n}(z-z_{o})^{n}.}
Using the integral formula for
f
(
n
)
{\displaystyle f^{(n)}}
, it immediately follows that the coefficients
a
n
{\displaystyle a_{n}}
are exactly the Taylor coefficients .
Estimation of the Taylor Series Coefficients
edit
For the coefficients, the following estimate holds when
|
f
(
z
)
|
≤
M
{\displaystyle |f(z)|\leq M}
for
|
z
−
a
|
<
r
⇔
z
∈
U
r
(
a
)
{\displaystyle |z-a|<r\ \Leftrightarrow z\in U_{r}(a)}
: :
|
a
n
|
≤
M
r
n
{\displaystyle |a_{n}|\leq {\frac {M}{r^{n}}}}
The Liouville Theorem (every holomorphic function bounded on the entire complex plane is constant) can be easily proven using the integral formula. This can then be used to easily prove the Fundamental Theorem of Algebra (every polynomial in
C
{\displaystyle \mathbb {C} }
factors into linear factors).
Here's the translation with the specified conditions:
The Cauchy integral formula is differentiated partially, allowing differentiation and integration to be swapped:
f
(
n
)
|
U
(
z
)
=
∂
n
f
∂
z
n
|
U
(
z
)
=
1
2
π
i
∂
n
∂
z
n
∮
∂
U
f
(
ζ
)
ζ
−
z
d
ζ
=
1
2
π
i
∮
∂
U
f
(
ζ
)
∂
n
∂
z
n
1
ζ
−
z
⏟
n
!
/
(
ζ
−
z
)
1
+
n
d
ζ
=
n
!
2
π
i
∮
∂
U
f
(
ζ
)
(
ζ
−
z
)
1
+
n
d
ζ
{\displaystyle {\begin{aligned}f^{(n)}|_{U}(z)&={\frac {\partial ^{n}f}{\partial z^{n}}}|_{U}(z)={\frac {1}{2\pi \mathrm {i} }}{\frac {\partial ^{n}}{\partial z^{n}}}\oint _{\partial U}{\frac {f(\zeta )}{\zeta -z}}\mathrm {d} \zeta \\&={\frac {1}{2\pi \mathrm {i} }}\oint _{\partial U}f(\zeta )\underbrace {{\frac {\partial ^{n}}{\partial z^{n}}}{\frac {1}{\zeta -z}}} _{n!/(\zeta -z)^{1+n}}\mathrm {d} \zeta ={\frac {n!}{2\pi \mathrm {i} }}\oint _{\partial U}{\frac {f(\zeta )}{(\zeta -z)^{1+n}}}\mathrm {d} \zeta \end{aligned}}}
Proof 2a: Cauchy Kernel
edit
Developing
1
ζ
−
z
{\displaystyle {\frac {1}{\zeta -z}}}
in the Cauchy integral formula using the geometric series gives (Cauchy kernel):
1
1
−
z
−
z
o
ζ
−
z
o
=
∑
n
=
0
∞
(
z
−
z
o
ζ
−
z
o
)
n
{\displaystyle {\frac {1}{1-{\frac {z-z_{o}}{\zeta -z_{o}}}}}=\sum _{n=0}^{\infty }\left({\frac {z-z_{o}}{\zeta -z_{o}}}\right)^{n}}
Proof 2: Cauchy Kernel - Taylor Series
edit
f
|
U
(
z
)
=
1
2
π
i
∮
∂
D
r
(
z
o
)
f
(
ζ
)
ζ
−
z
d
ζ
=
1
2
π
i
∮
∂
D
r
(
z
o
)
f
(
ζ
)
ζ
−
z
o
−
(
z
−
z
o
)
d
ζ
=
1
2
π
i
∮
∂
D
r
(
z
o
)
f
(
ζ
)
ζ
−
z
o
⋅
1
1
−
z
−
z
o
ζ
−
z
o
d
ζ
=
|
z
−
z
o
ζ
−
z
o
|
<
1
1
2
π
i
∮
∂
D
r
(
z
o
)
f
(
ζ
)
ζ
−
z
o
∑
n
=
0
∞
(
z
−
z
o
ζ
−
z
o
)
n
d
ζ
=
∑
n
=
0
∞
(
1
2
π
i
∮
∂
D
r
(
z
o
)
f
(
ζ
)
(
ζ
−
z
o
)
n
+
1
d
ζ
)
⏟
a
n
(
z
−
z
o
)
n
{\displaystyle {\begin{aligned}f|_{U}(z)&={\frac {1}{2\pi \mathrm {i} }}\oint _{\partial D_{r}(z_{o})}{\frac {f(\zeta )}{\zeta -z}}\mathrm {d} \zeta ={\frac {1}{2\pi \mathrm {i} }}\oint _{\partial D_{r}(z_{o})}{\frac {f(\zeta )}{\zeta -z_{o}-(z-z_{o})}}\mathrm {d} \zeta \\&{=}{\frac {1}{2\pi \mathrm {i} }}\oint _{\partial D_{r}(z_{o})}{\frac {f(\zeta )}{\zeta -z_{o}}}\cdot {\frac {1}{1-{\frac {z-z_{o}}{\zeta -z_{o}}}}}\mathrm {d} \zeta \,\\&{\overset {|{\frac {z-z_{o}}{\zeta -z_{o}}}|<1}{=}}{\frac {1}{2\pi \mathrm {i} }}\oint _{\partial D_{r}(z_{o})}{\frac {f(\zeta )}{\zeta -z_{o}}}\sum _{n=0}^{\infty }\left({\frac {z-z_{o}}{\zeta -z_{o}}}\right)^{n}\mathrm {d} \zeta \\&=\sum _{n=0}^{\infty }\underbrace {\left({\frac {1}{2\pi \mathrm {i} }}\oint _{\partial D_{r}(z_{o})}{\frac {f(\zeta )}{(\zeta -z_{o})^{n+1}}}\mathrm {d} \zeta \right)} _{a_{n}}(z-z_{o})^{n}\end{aligned}}}
Proof 2b: Cauchy Kernel
edit
Since the geometric series converges uniformly for
|
z
−
z
o
|
<
|
ζ
−
z
o
|
=
r
{\displaystyle |z-z_{o}|<|\zeta -z_{o}|=r}
, one can integrate term by term, i.e., swap the sum and the integral. The development coefficients are:
a
n
=
1
n
!
f
(
n
)
|
U
(
z
o
)
=
1
2
π
i
∮
∂
U
r
(
a
)
f
(
ζ
)
(
ζ
−
z
o
)
n
+
1
d
ζ
=
1
2
π
i
∫
0
2
π
f
(
z
o
+
r
e
i
t
)
(
r
e
i
t
)
n
+
1
i
r
e
i
t
d
t
=
1
2
π
r
n
∫
0
2
π
f
(
z
o
+
r
e
i
t
)
e
−
i
n
t
d
t
{\displaystyle {\begin{aligned}a_{n}&={\frac {1}{n!}}f^{(n)}|_{U}(z_{o})={\frac {1}{2\pi \mathrm {i} }}\oint _{\partial U_{r}(a)}{\frac {f(\zeta )}{(\zeta -z_{o})^{n+1}}}\mathrm {d} \zeta \\&={\frac {1}{2\pi \mathrm {i} }}\int _{0}^{2\pi }{\frac {f(z_{o}+re^{\mathrm {i} t})}{(re^{\mathrm {i} t})^{n+1}}}\mathrm {i} re^{\mathrm {i} t}\,\mathrm {d} t={\frac {1}{2\pi r^{n}}}\int _{0}^{2\pi }f(z_{o}+re^{\mathrm {i} t})e^{-\mathrm {i} nt}\,\mathrm {d} t\end{aligned}}}
Proof 3: Estimation of the Coefficients
edit
For the coefficients
a
n
∈
C
{\displaystyle a_{n}\in \mathbb {C} }
, the following estimate holds. There exists a
M
>
0
{\displaystyle M>0}
such that
|
f
(
z
)
|
≤
M
{\displaystyle |f(z)|\leq M}
for
|
z
−
z
o
|
=
r
{\displaystyle |z-z_{o}|=r}
. Then, for
n
∈
N
0
{\displaystyle n\in \mathbb {N} 0}
, we have:
|
a
n
|
=
|
1
2
π
r
n
∫
0
2
π
f
(
z
o
+
r
e
i
t
)
e
−
i
n
t
d
t
|
≤
1
2
π
r
n
∫
0
2
π
|
f
(
z
o
+
r
e
i
t
)
|
⏟
≤
M
d
t
≤
M
r
n
{\displaystyle {\begin{aligned}|a_{n}|&=\left|{\frac {1}{2\pi r^{n}}}\int _{0}^{2\pi }f(z_{o}+re^{\mathrm {i} t})e^{-\mathrm {i} nt}\,\mathrm {d} t\right|\\&\leq {\frac {1}{2\pi r^{n}}}\int _{0}^{2\pi }\underbrace {|f(z_{o}+re^{\mathrm {i} t})|} _{\leq M}\,\mathrm {d} t\leq {\frac {M}{r^{n}}}\end{aligned}}}
Proof 4: Liouville's Theorem
edit
If
f
{\displaystyle f}
is holomorphic on all of
C
{\displaystyle \mathbb {C} }
and bounded, i.e.,
|
f
(
z
)
|
=
|
∑
n
=
0
∞
a
n
z
n
|
≤
M
{\displaystyle |f(z)|=|\sum _{n=0}^{\infty }a_{n}z^{n}|\leq M}
for all
z
∈
C
{\displaystyle z\in \mathbb {C} }
, then, as before, for all
r
>
0
{\displaystyle r>0}
, we have:
|
a
n
|
≤
M
r
n
{\displaystyle |a_{n}|\leq {\frac {M}{r^{n}}}}
Since
r
{\displaystyle r}
was arbitrary, it follows that
a
n
=
0
{\displaystyle a_{n}=0}
for all
n
∈
N
{\displaystyle n\in \mathbb {N} }
. Therefore, from the boundedness of
f
{\displaystyle f}
, we conclude:
f
(
z
)
=
a
0
{\displaystyle f(z)=a_{0}}
Thus, every bounded holomorphic function on all of
C
{\displaystyle \mathbb {C} }
is constant (Liouville's theorem).
Using the integral formula, integrals can also be computed:
∮
∂
U
2
(
0
)
e
2
ζ
(
ζ
+
1
)
4
d
ζ
=
2
π
i
3
!
d
3
d
z
3
e
2
z
|
z
=
−
1
=
8
π
i
3
e
2
{\displaystyle \oint _{\partial U_{2}(0)}{\frac {e^{2\zeta }}{(\zeta +1)^{4}}}\mathrm {d} \zeta ={\frac {2\pi \mathrm {i} }{3!}}{\frac {\mathrm {d} ^{3}}{\mathrm {d} z^{3}}}e^{2z}|_{z=-1}={\frac {8\pi \mathrm {i} }{3e^{2}}}}
The Cauchy integral formula has been generalized to the multidimensional complex space
C
n
{\displaystyle \mathbb {C} ^{n}}
. Let
U
1
,
…
,
U
n
{\displaystyle U_{1},\ldots ,U_{n}}
be disk domains in
C
{\displaystyle \mathbb {C} }
, then
U
:=
∏
i
=
1
n
U
i
{\displaystyle U:=\prod _{i=1}^{n}U_{i}}
is a Polycylinder in
C
n
{\displaystyle \mathbb {C} ^{n}}
. Let
f
:
U
→
C
{\displaystyle f\colon U\to \mathbb {C} }
be a holomorphic function and
ξ
∈
U
{\displaystyle \xi \in U}
. The Cauchy integral formula is given by
f
(
z
1
,
…
,
z
n
)
=
1
(
2
π
i
)
n
∮
∂
U
n
⋯
∮
∂
U
1
f
(
ξ
1
,
…
,
ξ
n
)
(
ξ
1
−
z
1
)
⋯
(
ξ
n
−
z
n
)
d
ξ
1
⋯
d
ξ
n
{\displaystyle f(z_{1},\ldots ,z_{n})={\frac {1}{(2\pi \mathrm {i} )^{n}}}\oint _{\partial U_{n}}\cdots \oint _{\partial U_{1}}{\frac {f(\xi _{1},\ldots ,\xi _{n})}{(\xi _{1}-z_{1})\cdots (\xi _{n}-z_{n})}}\mathrm {d} \xi _{1}\cdots \mathrm {d} \xi _{n}}
Restrictions in Multidimensional Space
edit
Since the Cauchy integral theorem does not hold in higher-dimensional space, this formula cannot be derived analogously to the one-dimensional case. Therefore, this integral formula is derived using induction from the Cauchy integral formula for disk domains. Using the multi-index notation, the formula can be simplified to:
f
(
z
)
=
1
(
2
π
i
)
n
∮
∂
U
f
(
ξ
)
(
ξ
−
z
)
,
d
ξ
{\displaystyle f(z)={\frac {1}{(2\pi \mathrm {i} )^{n}}}\oint _{\partial U}{\frac {f(\xi )}{(\xi -z)}},\mathrm {d} \xi }
with
∂
U
=
∂
U
1
×
⋯
×
∂
U
n
{\displaystyle \partial U=\partial U_{1}\times \cdots \times \partial U_{n}}
.
Polycycles are defined using a vector of radii, where
M
:=
max
ξ
∈
U
|
f
(
ξ
)
|
{\displaystyle M:=\max _{\xi \in U}|f(\xi )|}
and
r
=
(
r
1
,
…
,
r
n
)
{\displaystyle r=(r_{1},\ldots ,r_{n})}
is the radius of the polycycle
U
:=
∏
i
=
1
n
U
i
{\displaystyle U:=\prod _{i=1}^{n}U_{i}}
.<ref>
for the derivatives of the holomorphic Function
f
{\displaystyle f}
as well as Cauchy's inequality
|
D
k
f
(
z
)
|
≤
M
⋅
k
!
r
k
,
{\displaystyle \left|D^{k}f(z)\right|\leq {\frac {M\cdot k!}{r^{k}}},}
Kurt Endl, Wolfgang Luh : Analysis. Volume 3: Function Theory, Differential Equations. 6th revised edition. Aula-Verlag, Wiesbaden 1987, ISBN 3-89104-456-9 , p. 153, Theorem 4.9.1.
Wolfgang Fischer, Ingo Lieb : Function Theory. 7th improved edition. Vieweg, Braunschweig, 1994, ISBN 3-528-67247-1 , p. 60, Chapter 3, Theorem 2.2 (Vieweg-Studium. Advanced Mathematics Course 47).