In [1]:
using Plots, ComplexPhasePortrait, ApproxFun
Dr. Sheehan Olver
s.olver@imperial.ac.uk
References:
Definition (Complex-differentiable) Let $D \subset {\mathbb C}$ be an open set. A function $f : D \rightarrow {\mathbb C}$ is called complex-differentiable at a point $z_0 \in D$ if $$ f'(z_0) = \lim_{z \rightarrow z_0} {f(z) - f(z_0) \over z - z_0} $$ exists, for any angle of approach to $z_0$.
Definition (Holomorphic) Let $D \subset {\mathbb C}$ be an open set. A function $f : D \rightarrow {\mathbb C}$ is called holomorphic in $D$ if it is complex-differentiable at all $z \in D$.
Definition (Entire) A function is entire if it is holomorphic in ${\mathbb C}$
Examples
We can usually infer the domain where a function is holomorphic from a phase portrait, here we see that ${\rm arcsinh}\, z$ has cuts on $[\I,\I \infty)$ and $[-\I,-\I \infty)$, and a zero (red–green–blue–red) at zero, hence we can infer that it is holomorphic in $\C \backslash ([\I,\I \infty) \cup [-\I,-\I \infty))$.
In [15]:
phaseplot(-4..4, -4..4, z -> asinh(z))
Out[15]:
The following example $\sqrt{z-1} \sqrt{z+1}$ is analytic in ${\mathbb C}\backslash [-1,1]$ and will be returned to:
In [3]:
phaseplot(-4..4, -4..4, z -> sqrt(z-1)sqrt(z+1))
Out[3]:
Definition (Contour) A contour is a continuous & piecewise-continuously differentiable function $\gamma : [a,b] \rightarrow {\mathbb C}$.
Definition (Simple) A simple contour is a contour that is 1-to-1.
Definition (Closed) A closed contour is a contour such that $\gamma(a) = \gamma(b)$
Examples of contours
Here's an example of a closed contour that is not simple:
In [13]:
a,b = -π, π
tt = range(a, stop=b, length=1000)
γ = t -> exp(im*t) +exp(2im*t)
plot(real.(γ.(tt)), imag.(γ.(tt)); ratio=1.0, legend=false, arrow=true)
Out[13]:
In [14]:
f = Fun( z -> real(exp(z)), Arc(0.,1.,(0,π/2))) # Not holomorphic!
plot(domain(f); legend=false, ratio=1.0, arrow=true)
Out[14]:
In [6]:
sum(f) # this means contour integral
Out[6]:
In [7]:
g = im*Fun(t-> f(exp(im*t))*exp(im*t), 0 .. π/2)
sum( g ) # this is standard integral
Out[7]:
An important property of a contour is its arclength:
Definition (Arclength) The arclength of $\gamma$ is defined as $$ {\cal L}(\gamma) := \int_a^b |\gamma'(t)| dt $$
A very useful result is that we can use the maximum and arclength to bound integrals:
Proposition (ML) Let $f : \gamma \rightarrow {\mathbb C}$ and
$$
M = \sup_{z \in \gamma} |f(z)|
$$
Then
$$
\left|\int_\gamma f(z) dz \right| \leq M {\cal L}(\gamma)
$$
In [8]:
f = Fun( z -> exp(z), Arc(0.,1.,(0,π/2))) # Not holomorphic!
sum(f) , f(im)-f(1)
Out[8]:
In [9]:
f = Fun( z -> exp(z), Arc(0.,1.,(0,π/2))) # Holomorphic!
sum(f) , f(im)-f(1)
Out[9]:
In [10]:
f = Fun( z -> exp(z), Segment(1,im)) # Holomorphic!
sum(f) , f(im)-f(1)
Out[10]:
Theorem (Cauchy) If $f$ is holomorphic inside and on a closed contour $\gamma$, then $$\oint_\gamma f(z) dz = 0$$
In [11]:
f = Fun( z -> real(exp(z)), Circle()) # Not holomorphic!
sum(f)
Out[11]:
In [12]:
f = Fun( z -> exp(z), Circle()) # Holomorphic!
sum(f)
Out[12]: