Sie sind auf Seite 1von 4

COMPLEX ANALYSIS

Real vs. Complex Differentiability.

1 A Sort of Preface
In this course we will not distinguish many times between a complex number z
and the pair of real numbers (x, y) such that z = x + iy. Mathematical objects
are basically defined by their properties and, as far as I am concerned, the “duck
principle” applies: If it walks like a duck, quacks like a duck, it is a duck. Or,
as Norbert Wiener put it in one of his books: (and I paraphrase here!) “Most
people will say that an egg is the device by which a chicken produces another
chicken. One might as well say that a chicken is the device by which an egg
produces another egg.” In other words, there is really no difference between a
chicken and an egg since both carry the same amount of information. To return
to the topic at hand, what this amounts to is that given a function f : Ω → C,
where Ω is a (usually open) subset of C, we will mostly write w = f (z), but
might suddenly switch to w = f (x, y) where (x, y) = x + iy ∈ Ω; x, y real.Or
to w = (u, v), u = u(x, y), v = v(x, y). To put it as briefly as I can, a complex
number is a pair of real numbers, and sometimes it is convenient to see it that
way. At a very basic level a complex valued function of a complex variable is
just a pair of real valued functions of two real variables. But a complex number
is also much more than just a pair of real numbers, just as chickens are much
more than eggs.

2 Real Differentiability
The definition I am about to give actually makes sense for functions with domain
in Rn , values in Rm , arbitrary n, m. But I will specialize to n = 2 and m = 1
or 2. Let Ω be an open subset of R2 = C and let g : Ω → R. We say g is
differentiable at a point z0 = (x0 , y0 ) ∈ Ω iff there exist real numbers a, b such
that
|g(x − x0 , y − y0 ) − g(x0 , y0 ) − a(x − x0 ) − b(y − y0 )|
lim √ = 0.
(x,y)→(x0 ,y0 ) (x − x0 )2 + (y − y0 )2

A more elegant definition talks of linear mappings, but this will do for now. In
terms of ϵ, δ notation, the definition is: g is differentiable at z0 = (x0 , y0 ) iff
there exist real numbers a, b such that for every ϵ > 0 there is δ > 0 with the
property that

|g(x − x0 , y − y0 ) − g(x0 , y0 ) − a(x − x0 ) − b(y − y0 )| ≤ ϵ (x − x0 )2 + (y − y0 )2

whenever (x, y) ∈ Ω and (x − x0 )2 + (y − y0 )2 < δ. One can compactify the
notation by using complex numbers, writing z = (x, y); then it becomes: g is

1
differentiable at z0 iff there exist real numbers a, b such that for every ϵ > 0
there is δ > 0 with the property that

|g(z) − g(z0 ) − aRe(z − z0 ) − bIm(y − y0 )| ≤ ϵ|z − z0 |

whenever z ∈ Ω and |z − z0 | < δ.


One now has the following facts, most (probably all?) of which are proved
in standard Calculus textbooks.
1. If g is differentiable at z0 = (x0 , y0 ) then it follows that both partial
derivatives of g at (x0 , y0 ) exist and, in fact, one must have

∂g ∂g
a= (x0 , y0 ), b= (x0 , y0 ).
∂x ∂y

2. If g is differentiable at z0 , then g is continuous at z0 .


3. Simple examples show that a function may have both partials at a point
without even being continuous at that point. This is not so strange since
the partials only govern the behavior of the function along lines paral-
lel to the x and y axes. The situation is different if the function has
partial derivatives in a whole neighborhood of the point. A basic result
∂g ∂g
is: If , exist at all points of an open disk centered at z0 , and are
∂x ∂y
continuous at z0 , then g is differentiable at z0 .

Suppose now Ω is again an open subset of R2 = C, z0 ∈ Ω and f : Ω → C.


We can then write

f (z) = u(z) + iv(z), or f (x, y) = (u(x, y), v(x, y))

where u = Ref, v = Imf . We say f is differentiable at z0 iff both u and v are


differentiable at z0 . We then define
∂f ∂u ∂v ∂f ∂u ∂v
= +i , = +i .
∂x ∂x ∂x ∂y ∂y ∂y
It is a simple exercise to verify that f : Ω → C is differentiable at z0 iff there
exist complex numbers α = a + ic, β = b + id such that

|f (z) − f (z0 ) − α(x − x0 ) − β(y − y0 )|


(1) lim = 0,
z→z0 |z − z0 |

where z = (x, y). In this case

∂f ∂f
α= (z0 ), β= (z0 ).
∂x ∂y

2
3 Complex Differentiability
Once again let Ω be an open subset of C and z0 ∈ Ω. We say f : Ω → C is
complex differentiable at z0 iff

f (z) − f (z0 )
lim
z→z0 z − z0
exists; equivalently iff there exists a complex number α such that

f (z) − f (z0 )
(2) lim − α = 0.
z→z0 z − z0

A big difference between (1) and (2) is that in (2) we are dividing by z − z0 ; we
are using the fact that C is a field. In (1) we are only dividing by real numbers.
I mention here a difference of opinion with the textbook of Stein and Shakarchi:
The textbook uses in a completely interchangeable way “complex differentiable”
and “holomorphic.” I will stick to the more traditional way and define a function
f : Ω → C to be holomorphic or analytic at z0 ∈ Ω iff there is a neighborhood V
of z0 in Ω such that f is complex differentiable at all points of V . Equivalently,
iff there exists r > 0 such that Dr (z0 ) ⊂ Ω and f is complex differentiable at
all points of Dr (z0 ). You can verify as an exercise that with this definition the
function f (z) = |z|2 is complex differentiable at 0, but holomorphic nowhere.
As it turns out, complex differentiability is a more potent concept than
real differentiability. The relation between real and complex differentiability is
contained in the following theorem.

Theorem 1 Let Ω be an open subset of C, let f : Ω → C, let z0 ∈ Ω. Then f


is complex differentiable at z0 if and only if f is differentiable at z0 (in the real
sense) and

∂f 1 ∂f
(3) (z0 ) = (z0 ).
∂x i ∂y
∂f
Moreover, if f is complex differentiable at z0 , then f ′ (z0 ) = (z0 ).
∂x
Proof. Assume first f is complex differentiable at z0 . Then

|f (z) − f (z0 ) − f ′ (z0 )(z − z0 )| f (z) − f (z0 )
(4) lim
= lim − f (z0 ) = 0.

z→z0 |z − z0 | z→z0 z − z0

If we notice that

f ′ (z0 )(z − z0 ) = f ′ (z0 ) (Re(z − z0 ) + iIm(z − z0 )) ,

comparing (4) with (1), we see that (1) holds with α = f ′ (z0 ), β = if ′ (z0 ).

3
Conversely, assume that f is (real) differentiable at z0 and (3) holds. Then
(1) holds with β = iα so that

|f (z) − f (z0 ) − α(x − x0 ) − iα(y − y0 )|


lim = 0;
z→z0 |z − z0 |

i.e.,
|f (z) − f (z0 ) − α(z − z0 )|
lim = 0.
z→z0 |z − z0 |
This is equivalent to
f (z) − f (z0 )
lim − α = 0,
z→z0 z − z0
∂f
hence f is complex differentiable and (z0 ) = α = f ′ (z0 ).
∂x
Equation (3) is usually written broken up into real and imaginary parts; that
is, as
∂u ∂v ∂u ∂v
(5) (z0 ) = (z0 ), (z0 ) = − (z0 )
∂x ∂y ∂y ∂x

if f = u + iv. One refers to the equations (5) as the Cauchy-Riemann equations.

Das könnte Ihnen auch gefallen