631019.txt

An Introduction to Complex Differentials and
Complex Differentiability

Raphael Hunger

Technical Report
TUM-LNS-TR-07-06
2007

Technische Universit¨ t M¨ nchen
a u
Associate Institute for Signal Processing
Prof. Dr.-Ing. Wolfgang Utschick

Contents

1. Introduction

3

2. Complex Differentiability and Holomorphic Functions

4

3. Differentials of Analytic and Non-Analytic Functions

8

4. Differentials of Real-Valued Functions

11

5. Derivatives of Functions of Several Complex Variables

14

6. Matrix-Valued Derivatives of Real-Valued Scalar-Fields

17

Bibliography

20

2

1. Introduction

This technical report gives a brief introduction to some elements of complex function theory. First,
general definitions for complex differentiability and holomorphic functions are presented. Since
non-analytic functions are not complex differentiable, the concept of differentials is explained both
for complex-valued and real-valued mappings. Finally, multivariate differentials and Wirtinger
derivatives are investigated.

3

2. Complex Differentiability and Holomorphic Functions

Complex differentiability is defined as follows, cf. [Schmieder, 1993, Palka, 1991]:
Definition 2.0.1. Let A ⊂ C be an open set. The function f : A → C is said to be (complex)
differentiable at z0 ∈ A if the limit
f (z) − f (z0 )
lim
(2.1)
z→z0
z − z0
exists independent of the manner in which z → z0 . This limit is then denoted by f (z0 ) =
and is called the derivative of f with respect to z at the point z0 .

df (z)
dz z=z0

A similar expression for (2.1) known from real analysis reads as
df (z)
f (z + ∆z) − f (z)
= lim
,
∆z→0
dz
∆z

(2.2)

where ∆z ∈ C now holds. Note that if f is differentiable at z0 then f is continuous at z0 . An
equivalent, but geometrically more illuminating way to define the derivative follows from the linear
approximation of f in the local vicinity of z0 [Palka, 1991].
Definition 2.0.2. Let A be an open set. The function f : A → C is said to be (complex) differentiable at z0 ∈ A if there exists a complex-valued scalar g such that
f (z) = f (z0 ) + g · (z − z0 ) + e(z, z0 ),

(2.3)

holds for every z ∈ A and the function e(·, ·) satisfies the condition
lim

z→z0

e(z, z0 )
= 0.
z − z0

4

(2.4)

2. Complex Differentiability and Holomorphic Functions

5

The remainder term e(z, z0 ) in (2.4) obviously is o(|z − z0 |) for z → z0 and therefore g · (z − z0 )
dominates e(z, z0 ) in the immediate vicinity of z0 if g = 0. Close to z0 , the differentiable function
f (z) can linearly be approximated by f (z0 ) + f (z0 )(z − z0 ). The difference z − z0 is rotated by
∠f (z0 ), scaled by |f (z0 )| and afterwards shifted by f (z0 ).
The concept of a differentiability in a single point readily extends to differentiability in open sets.

Definition 2.0.3. Let U ⊆ A be a nonempty open set. The function f : A → C is called holomorphic (or analytic) in U, if f is differentiable in z0 for all z0 ∈ U. Moreover, if f is analytic in the
complete open domain-set A, f is a holomorphic (analytic) function.

An interesting characteristic of a function f analytic in U is the fact that its derivative f is analytic
in U itself [Spiegel, 1974]. By induction, it can be shown that derivatives of all orders exist and
are analytic in U which is in contrast to real-valued functions, where continuous derivatives need
not be differentiable in general. However, basic properties for the derivative of a sum, product, and
composition of two functions known from real-valued analysis remain inherently valid in the complex domain. Assume that f (z) and g(z) are differentiable at z0 . Then, the following propositions
hold:
Proposition 2.0.1. The sum f + g is differentiable at z0 and
(f + g) (z0 ) = f (z0 ) + g (z0 ).

(2.5)

Proposition 2.0.2. The product f g is differentiable at z0 and
(f g) (z0 ) = f (z0 )g(z0 ) + f (z0 )g (z0 ).

Proposition 2.0.3. If g(z0 ) = 0, the quotient
f
g

(z0 ) =

f
g

(2.6)

is differentiable at z0 and

f (z0 )g(z0 ) − f (z0 )g (z0 )
.
g 2 (z0 )

(2.7)

Proposition 2.0.4. If f is differentiable at g(z0 ), the composition f ◦ g is differentiable at z0 and
(f ◦ g) (z0 ) = f (g(z0 ))g (z0 ) (chain rule).

(2.8)

Complex differentiability is closely related to the Cauchy-Riemann equations [Lang, 1993]. A necessary condition for f being holomorphic in U requires the Cauchy-Riemann equations to be satisfied.

6

2. Complex Differentiability and Holomorphic Functions

Theorem 2.0.1: Let f (z) = u(z) + jv(z) with u(z), v(z) ∈ R and z = x + jy with x, y ∈ R.
In terms of x and y, the function f (z) can be expressed as F (x, y) = U (x, y) + jV (x, y) with
U (x, y), V (x, y) ∈ R. A necessary condition for f (z) being holomorphic in U is that the following
system of partial differential equations termed Cauchy-Riemann-equations holds for every z =
x + jy ∈ U:
∂V (x, y)
∂U (x, y)
∂V (x, y)
∂U (x, y)
=
and
=−
.
(2.9)
∂x
∂y
∂y
∂x

Proof : According to Definition 2.0.3, f (z) is holomorphic in U if f (z) is differentiable at every
z ∈ U. Differentiability at z implies that the limit
f (z + ∆z) − f (z)
∆z→0
∆z

= lim

lim

z=x+jy

∆x→0
∆y→0

F (x + ∆x, y + ∆y) − F (x, y)
∆x + j∆y

exists no matter which curve ∆z moves along when approaching zero, see Definition 2.0.1
and (2.2). Setting ∆z = ∆x + j∆y, two possible curves for ∆z → 0 are considered. The first
curve goes in the horizontal direction with ∆y = 0 and ∆x → 0 yielding
F (x + ∆x, y) − F (x, y)
∆x→0
∆x
U (x + ∆x, y) − U (x, y)
V (x + ∆x, y) − V (x, y)
+j
= lim
∆x→0
∆x
∆x
∂V (x, y)
∂U (x, y)
+j
.
=
∂x
∂x
The second curve goes in the vertical direction with ∆x = 0 and ∆y → 0 yielding
f (z = x + jy) = lim

F (x, y + ∆y) − F (x, y)
∆y→0
j∆y
V (x, y + ∆y) − V (x, y)
U (x, y + ∆y) − U (x, y)
+j
= lim
∆y→0
j∆y
j∆y
∂U (x, y) ∂V (x, y)
+
.
=
j∂y
∂y
As both expressions have to be the same for f (z) being holomorphic, (2.9) immediately follows
as a necessary condition.
f (z = x + jy) = lim

2
The next theorem provides conditions under which the Cauchy-Riemann equations are sufficient
for f (z) being holomorphic.
Theorem 2.0.2: If the partial derivatives of U (x, y) and V (x, y) with respect to x and y are continuous, the Cauchy-Riemann equations are sufficient for f (z) being holomorphic.

Proof :

See [Spiegel, 1974].

2

2. Complex Differentiability and Holomorphic Functions
In the following, we give examples for analytic functions and functions which are not analytic.
Examples for analytic functions:
•

f (z) = z n

•

f (z) =

•

f (z) = ln(z)

•

f (z) = exp(az)

az + b
cz + d

f (z) = nz n−1
ad − bc
(cz + d)2
1
f (z) =
z
f (z) =

f (z) = a exp(az)

Examples for non-analytic functions:
• f (z) = |z|2
• f (z) = {z}
• f (z) = {z}
• f (z) = z ∗

7

3. Differentials of Analytic and Non-Analytic Functions

The total differential of the bivariate function F (x, y) associated to the univariate function f (z)
via F (x, y) = U (x, y) + jV (x, y) = f (z)|z=x+jy reads as [Henrici, 1974]
dF =

∂F (x, y)
∂F (x, y)
dx +
dy.
∂x
∂y

(3.1)

Of course, differentiability of F (x, y) with respect to x and y in the real sense has to be imposed
for the existence of the differential dF in (3.1). This implies the differentiability of the real-valued
functions U (x, y) and V (x, y) with respect to x and y. Rewriting (3.1) by means of F (x, y) =
U (x, y) + jV (x, y) yields
dF =

∂U (x, y)
∂V (x, y)
∂U (x, y)
∂V (x, y)
dx + j
dx +
dy + j
dy.
∂x
∂x
∂y
∂y

(3.2)

Making use of
dz = dx + jdy,
dz ∗ = dx − jdy,

(3.3)

the two differentials dx and dy can be expressed via
1
(dz + dz ∗ )
2
1
dy = (dz − dz ∗ ) .
2j

dx =

(3.4)

Inserting (3.4) into the differential expression dF in (3.1) and reordering the result leads to
dF =

1 ∂U (x, y) ∂V (x, y)
∂V (x, y) ∂U (x, y)
+
+j
−
dz
2
∂x
∂y
∂x
∂y
1 ∂U (x, y) ∂V (x, y)
∂V (x, y) ∂U (x, y)
+
dz ∗ .
−
+j
+
2
∂x
∂y
∂x
∂y

(3.5)

A major result can already be anticipated here.
Proposition 3.0.1. The differential of any analytical function f (z) does not depend on the differential dz ∗ .
8

3. Differentials of Analytic and Non-Analytic Functions

9

Proof : Since any analytical function f (z) satisfies the Cauchy-Riemann equations in (2.9), the
factor in front of dz ∗ in the second line of (3.5) is zero. Obviously, the differential dF does not
depend on dz ∗ .
2
Note that the converse of Proposition 3.0.1 is also true: If the differential of a function f does not
depend on dz ∗ , the function f is analytical.
Rearranging the terms in (3.5), the differential dF can be expressed as
dF =

∂
1 ∂
(U (x, y) + jV (x, y)) − j (U (x, y) + jV (x, y)) dz
2 ∂x
∂y
1 ∂
∂
+
(U (x, y) + jV (x, y)) + j (U (x, y) + jV (x, y)) dz ∗ .
2 ∂x
∂y

Recognizing that U (x, y) + jV (x, y) = F (x, y), we finally obtain by factoring out the partial
differential operators
dF =

∂
∂
1 ∂
1 ∂
−j
+j
F (x, y) dz +
F (x, y) dz ∗ .
2 ∂x
∂y
2 ∂x
∂y

(3.6)

According to the total differential for real-valued multivariate functions, the introduction of the
∂
∂
two operators ∂z and ∂z ∗ is reasonable as it leads to the very nice description of the differential df ,
where the real-valued partial derivatives are hidden [Trapp, 1996].
Theorem 3.0.1: The differential df of a complex-valued function f (z) : A → C with A ⊂ C
can be expressed as
∂f (z)
∂f (z) ∗
df =
dz .
(3.7)
dz +
∂z
∂z ∗

Proof : See the preceding derivation and the definition of the partial derivatives
following.

∂
∂z

and

∂
∂z ∗

in the
2

Definition 3.0.1. The two ‘partial derivative’ operators

∂
∂z

and

1 ∂
∂
∂
:=
−j
,
∂z
2 ∂x
∂y
∂
∂
1 ∂
+j
:=
,
∗
∂z
2 ∂x
∂y

∂
∂z ∗

are defined by

(3.8)

and are often referred to as the Wirtinger derivatives [Wirtinger, 1926] and correspond to half of
the del-bar and del operator [Spiegel, 1974].

10

3. Differentials of Analytic and Non-Analytic Functions

Basic rules for this Wirtinger calculus are given in the next theorem.
Theorem 3.0.2: For the Wirtinger derivatives, the common rules for differentiation known from
real-valued analysis concerning the sum, product, and composition of two functions hold as well.
In particular,
∂ ∗
∂
z = ∗ z = 0,
∂z
∂z
which means that z ∗ can be regarded as a constant when differentiating with respect to z, as well
as z can be regarded constant when differentiating with respect to z ∗ .

Proof :

With z = x + jy and z ∗ = x − jy, Theorem 3.0.2 follows immediately from (3.8).

2

Examples:
•

∂ 2
∂
|z| =
(zz ∗ ) = z ∗
∂z
∂z

•

∂
∂
exp −|z|2 =
exp (−zz ∗ ) = −z ∗ exp |z|2
∂z
∂z

Corollary 3.0.1. Derivatives of the conjugate function f ∗ (z) satisfy the relationships
∂f ∗ (z)
=
∂z

Proof :

∂f (z)
∂z ∗

∗

and

∂f ∗ (z)
=
∂z ∗

See the definition of the Wirtinger derivatives in (3.8).

∂f (z)
∂z

∗

.

(3.10)

2

4. Differentials of Real-Valued Functions

Optimizations in communications and signal processing are frequently targeted on the maximization of a utility or on the minimization of a cost. Hence, most objectives are real-valued, as the
standard total order can only handle real-valued arguments. On account of this, this chapter deals
with functions f (z) : U → R having complex-valued arguments z ∈ U ⊂ C that are mapped
to real-valued scalars f (z) ∈ R. In addition, simplifications resulting from this circumstance are
investigated.
First of all it is obvious that the only possibility of a real-valued function f (z) with complex
argument z for being analytic is that f (z) is constant for all z of its domain. This follows from the
Cauchy-Riemann equations in (2.9), since v(z) = V (x, y) = {f (z)} = 0 for real-valued f (z).
This leads to the following proposition.
Proposition 4.0.1. All non-trivial (not constant) real-valued functions f (z) mapping z ∈ A ⊂ C
onto R are non-analytic functions and therefore not complex differentiable.

With the definition of the differential dF in (3.2) and (3.6), it is easy to prove the following theorem.
Theorem 4.0.1: The differential df of a real-valued function f (z) : A → R with complex-valued
argument z ∈ A ⊂ C can be expressed as
df = 2

∂f (z)
dz
∂z

=2

∂f (z) ∗
dz
∂z ∗

(4.1)

and is equivalent to
dF =

∂F (x, y)
∂F (x, y)
dx +
dy.
∂x
∂y

(4.2)

Due to the property that the non-trivial real-valued functions are not analytic, stationary points of
the real-valued f (z) cannot be obtained by searching for points z where the derivative f (z) is
zero. However, we can detect stationary points z of f (z) by a vanishing differential df .
11

12

4. Differentials of Real-Valued Functions

Theorem 4.0.2: The differential df of a real-valued function f (z) : A → R with complex
argument z ∈ A ⊂ C vanishes if and only if the Wirtinger derivative is zero:
df = 0 ⇔

∂f (z)
= 0.
∂z

(4.3)

Proof : First, we prove that ∂f (z) = 0 leads to df = 0. This is a result from (4.1). The converse is
∂z
shown by the following reasoning. For arbitrary ratios of dx and dy, dF from (4.2) and therefore
df can only vanish if both partial derivatives of F (x, y) with respect to x and y are zero. With
∂f (z)
1 ∂U (x, y)
∂U (x, y)
(dx + jdy) ,
dz =
−j
∂z
2
∂x
∂y

(4.4)

vanishing partial derivatives document the second way of the equivalence relation in Theorem 4.0.2
as F (x, y) = U (x, y) for real-valued f (z).
2
Gradient-based iterative algorithms targeted on maximizing or minimizing an objective function
can be constructed by optimizing dz in the differential expression (4.1).
Corollary 4.0.1. The steepest ascent of a real-valued function f (z) : A → R with complexvalued argument z ∈ A ⊂ C is obtained for
dz =

∂f (z)
ds,
∂z ∗

where ds is a real-valued differential. Thus, the steepest ascent points to the direction of

Proof :

(4.5)
∂f (z)
.
∂z ∗

As the differential df of a real-valued function can be expressed as (see Equ. 4.1)
df = 2

∂f (z) ∗
dz ,
∂z ∗

df is maximized for real-valued ∂f (z) dz ∗ if the norm of dz is fixed. Hence, dz ∗ has to be a scaled
∂z ∗
∂f (z)
version of the conjugate of ∂z ∗ . Equivalently, dz must be a scaled version of ∂f (z) and (4.5)
∂z ∗
immediately follows.
2
An iterative implementation could therefore read as
z ←z+2

∂f (z)
ds,
∂z ∗

4. Differentials of Real-Valued Functions

13

where ds can be interpreted as the step-size. Notice that there is a factor 2 in front of the Wirtinger
derivative which follows from (4.1)!

5. Derivatives of Functions of Several Complex Variables

When switching to functions of several complex variables stacked in the column vector z =
[z1 , . . . , zn ]T ∈ Cn , we confine ourselves to mappings onto the one-dimensional complex domain C. For them, the holomorphic-property is defined by the following two definitions which are
equivalent [Krantz, 1992]:
Definition 5.0.1. A function f (z) : Cn ⊃ A → C is said to be holomorphic if for each k ∈
{1, . . . , n} and each fixed z1 , . . . , zk−1 , zk+1 , . . . , zn the function
w → f ([z1 , . . . , zk−1 , w, zk+1 , . . . , zn ]T )
is holomorphic according to the one-dimensional sense in Definition 2.0.3 on the set {w ∈ C :
[z1 , . . . , zk−1 , w, zk+1 , . . . , zn ]T ∈ A}.

This is nothing else than that the function f (z) has to be holomorphic in each variable z 1 , . . . , zn .
An equivalent definition reads as follows [Krantz, 1992]:
Definition 5.0.2. A function f (z) : A → C that is continuously differentiable in each variable z k ,
k ∈ {1, . . . , n} is said to be holomorphic if the Cauchy-Riemann equations are satisfied in each
variable separately.

Although there are many differences between univariate and multivariate complex functions, the
Wirtinger calculus easily extends to the case of several complex variables.
Theorem 5.0.1: The differential df of a multivariate complex-valued function f (z) : A → C
with A ⊂ Cn can be expressed as
n

df =
k=1

∂f (z)
dzk +
∂zk

n

k=1

∂f (z) ∗
dzk
∗
∂zk

∂f (z)
∂f (z) ∗
=
dz +
dz .
T
∂z
∂z H

14

(5.2)

5. Derivatives of Functions of Several Complex Variables
Note that the operators

∂
∂z

and

∂
∂z ∗

15

read as
∂
∂
∂
=
,...,
∂z
∂z1
∂zn

T

∂
∂
∂
=
,..., ∗
∗
∗
∂z
∂z1
∂zn

T

,
(5.3)
.

If they are applied to a scalar field f (z) they generate a column vector of dimension n and mimic
the gradient operator for real-valued functions. Again, both the differential dz = [dz 1 , . . . , dzn ]T
and its conjugate dz ∗ are required to express the differential df in (5.2) for non-analytic functions.
Similar to Theorem 4.0.2, we make the following observation for mappings onto R:
Theorem 5.0.2: The differential df of a real-valued function f (z) : A → R with complex
argument z ∈ A ⊂ Cn vanishes if and only if the vector-valued Wirtinger derivative is zero:
∂f (z)
= 0.
∂z

df = 0 ⇔

(5.4)

Finally, the conjugate Wirtinger derivative again points into the direction of the steepest ascent:
Corollary 5.0.1. The steepest ascent of a real-valued function f (z) : A → R with complexvalued argument z ∈ A ⊂ Cn is obtained for
dz =

∂f (z)
ds,
∂z ∗

(5.5)

where ds is a real-valued differential.

Proof :

From Theorem 5.0.1, the differential of the function f with real-valued image reads as
df = 2

∂f (z)
dz .
∂z T

According to the Cauchy-Schwarz inequality, dz has to be the conjugate of
valued differential and the proof is complete.

∂f (z) T
∂z T

times a real2

A gradient ascent step could for example read as
z ←z+2
with the step-size ds.

∂f (z)
ds
∂z ∗

(5.7)

16

5. Derivatives of Functions of Several Complex Variables

Examples for vector-valued Wirtinger derivatives:
∂f (z)
= AT z ∗
• f (z) = z H Az
∂z
∂f (z)
= zHA
∂z T
∂f (z)
= Az
∂z ∗
∂f (z)
= z T AT
H
∂z

6. Matrix-Valued Derivatives of Real-Valued Scalar-Fields

In this section, derivatives of real-valued scalar-fields are investigated. Common representatives of
such scalar fields in communications and signal processing are trace or determinant expressions.
Definition 6.0.1. Let f : Cn×n → R denote a functional acting as a map A → f (A). Then the
derivative of f (A) with respect to the matrix A returns a matrix-valued function whose entry in
the m-th row and n-th column reads as
∂
f (A)
∂A

=
m,n

∂
f (A).
∂[A]m,n

(6.1)

Sometimes the functional f is a composition of several operations where the outer one is a linear
operator like the trace-operator for example. In such a case, the partial derivative operator and this
outer linear operator may be interchanged as their commutator vanishes:
∂
∂
tr(A(t)) = tr
A(t) .
∂t
∂t
Clearly, derivatives of matrices with respect to scalars turn then out to be necessary and will therefore be discussed now.
The derivative of a matrix A(t) with respect to the variable t the matrix depends on follows from
the element-wise application of the partial derivative operator onto the entries of A. Hence, the
element in the m-th row and n-th column of the derivative reads as
∂
A(t)
∂t

=
m,n

∂
[A(t)]m,n .
∂t

(6.3)

If t is complex valued, then the partial derivative operator denotes the Wirtinger derivative. Equivalently, t may stand for an element of the matrix A(t). For example, if t = [A] m,n = am,n we
have
∂
A = e m eT ,
(6.4)
n
∂am,n
where em denotes the m-th canonical unit vector of appropriate dimension the elements of which
are all zero except for the one in the m-th row. For the following examples, all matrices are assumed
17

18

6. Matrix-Valued Derivatives of Real-Valued Scalar-Fields

to be constant and mutually independent. Moreover, no special structure or symmetry is assumed
for them.
Examples:
n
n
n
n
∂A
∂f (A)
T
=
=
• f (A) = tr(A)
ek e tr
ek eT tr ek eT
∂A
∂[A]k,
k=1 =1
k=1 =1
n

n

n

T T

=
k=1 =1

•

f (A) = tr(AB)

ek eT = I n
k

ek e e ek =

∂f (A)
=
∂A

n

k=1
n

∂AB
ek e tr
∂[A]k,
=1

n

T

k=1
n
n

ek eT tr ek eT B

=
n

n

k=1 =1

ek eT eT Bek =

=
k=1 =1

n

ek eT [B]l,k = B T
k=1 =1

Many information theoretic expressions involve the determinant-operator. For the derivative of
them, the following proposition holds:
Proposition 6.0.1. The derivative of the determinant of a matrix A(t) which depends on a parameter t with respect to this parameter reads as
∂A(t)
∂
.
det(A(t)) = det(A) tr A−1
∂t
∂t

∂
1
det(A(t)) = lim
[det(A(t + ∆t)) − det(A(t))]
∆t→0 ∆t
∂t
1
[det(A(t) + ∆tB(t)) − det(A(t))]
= lim
∆t→0 ∆t
1
= det(A(t)) lim
det(In + ∆tA−1 (t)B(t)) − 1 ,
∆t→0 ∆t

Proof :

(6.5)

(6.6)

We have

∂
where B(t) = ∂t A(t). Making use of the Schur-decomposition [Golub and Loan, 1991], we can
−1
rewrite A (t)B(t) to A−1 (t)B(t) = QΛQH with unitary Q and the upper triangular matrix Λ
the diagonal values of which are the eigenvalues λi , i ∈ {1, . . . , n}. We get

∂
1
det(A(t)) = det(A(t)) lim
∆t→0 ∆t
∂t

n

(1 + ∆tλi ) − 1
i=1

n

= det(A(t))

λi = det(A(t)) tr(Λ) =
i=1

= det(A(t)) tr A−1 (t)B(t) .
This completes the proof.
2
From differentiating the identity matrix In = A(t)A−1 (t) with respect to t by means of the product
rule, we obtain the following proposition:

6. Matrix-Valued Derivatives of Real-Valued Scalar-Fields

19

Proposition 6.0.2. The derivative of the inverse of a matrix A(t) with respect to the parameter t
reads as
∂A(t) −1
∂ −1
(6.7)
A (t) = −A−1 (t)
A (t).
∂t
∂t

Bibliography

[Golub and Loan, 1991] Golub, G. H. and Loan, C. V. (1991). Matrix Computations. Johns Hopkins University Press.
[Henrici, 1974] Henrici, P. (1974). Applied and Computational Complex Analysis (Volume I),
volume 1. Wiley-Interscience.
[Krantz, 1992] Krantz, S. G. (1992). Function Theory of Several Complex Variables. American
Mathematical Society.
[Lang, 1993] Lang, S. (1993). Complex Analysis (Graduate Texts in Mathematics). Springer.
[Palka, 1991] Palka, B. P. (1991). An Introduction to Complex Function Theory (Undergraduate
Texts in Mathematics). Springer.
[Schmieder, 1993] Schmieder, G. (1993). Grundkurs Funktionentheorie. Teubner Verlag.
[Spiegel, 1974] Spiegel, M. (1974). THEORY AND PROBLEMS OF Complex Variables with an
introduction to CONFORMAL MAPPING and its applications. McGraw Hill Book Company.
[Trapp, 1996] Trapp, H. (1996). Funktionentheorie einer Ver anderlichen. Universit¨
atsverlag Os¨
nabr¨ bei V&R unipress.
uck
[Wirtinger, 1926] Wirtinger, W. (1926). Zur formalen Theorie der Funktionen von mehr komanderlichen.
plexen Ver¨

20