4.4 KiB
Math416 Lecture 17
Continue on Chapter 7
Harmonic conjugates
Theorem 7.18
Existence of harmonic conjugates.
Let u be a harmonic function on \Omega a convex open subset in \mathbb{C}. Then there exists g\in O(\Omega) such that \text{Re}(g)=u on \Omega.
Moreover, g is unique up to an imaginary additive constant.
Proof:
Let f=2\frac{\partial u}{\partial z}=\frac{\partial u}{\partial x}-i\frac{\partial u}{\partial y}
f is holomorphic on \Omega
Since \frac{\partial u}{\partial \overline{z}}=0 on \Omega, f is holomorphic on \Omega
So f=g', fix z_0\in \Omega, we can choose q(z_0)=u(z_0) and g=u_1+iv_1, g'=\frac{\partial u_1}{\partial x}+i\frac{\partial v_1}{\partial x}=\frac{\partial v_1}{\partial y}-i\frac{\partial u_1}{\partial y}=\frac{\partial u}{\partial x}-i\frac{\partial u}{\partial y}, given that \frac{\partial u_1}{\partial x}=\frac{\partial u}{\partial x} and \frac{\partial u_1}{\partial y}=\frac{\partial u}{\partial y}
So u_1=u on \Omega
\text{Re}(g)=u_1=u on \Omega
If u+iv is holomorphic, v is harmonic conjugate of u
QED
Corollary For Harmonic functions
Theorem 7.19
Harmonic functions are C^\infty
C^\infty is a local property.
Theorem 7.20
Mean value property for harmonic functions.
Let u be harmonic on an open set of \Omega
Then u(z_0)=\frac{1}{2\pi}\int_0^{2\pi}u(z_0+re^{i\theta})d\theta
Proof:
\text{Re}g(z_0)=\frac{1}{2\pi}\int_0^{2\pi}\text{Re}g(z_0+re^{i\theta})d\theta
QED
Theorem 7.21
Identity theorem for harmonic functions.
Let u be harmonic on a domain \Omega. If u=0 on some open set G\subset \Omega, then u\equiv 0 on \Omega.
If u=v on G\subset \Omega, then u=v on \Omega.
Proof:
We proceed by contradiction.
Let H=\{z\in \Omega:u(z)=0\} be the interior of G
H is open and nonempty. If H\neq \Omega, then \exists z_0\in \partial H\cap \Omega. Then \exists r>0 such that B_r(z_0)\subset \Omega such that \exists g\in O(B_r(z_0)) such that \text{Re}g=u on B_r(z_0)
Since H\cap B_r(z_0) is nonempty open set, then g is constant on H\cap B_r(z_0)
So g is constant on B_r(z_0)
So u is constant on B_r(z_0)
So D(z_0,r)\subset H. This is a contradiction that z_0\in \partial H
QED
Theorem 7.22
Maximum principle for harmonic functions.
A non-constant harmonic function on a domain cannot attain a maximum or minimum on the interior of the domain.
Proof:
We proceed by contradiction.
Suppose u attains a maximum at z_0\in \Omega.
For all z in the neighborhood of z_0, u(z)<u(z_0). We can choose r>0 such that B_r(z_0)\subset \Omega.
By the mean value property, u(z_0)=\frac{1}{2\pi}\int_0^{2\pi}u(z_0+re^{i\theta})d\theta
So 0= \frac{1}{2\pi}\int_0^{2\pi}u[z_0+re^{i\theta}-u(z_0)]d\theta
We can prove the minimum is similar.
QED
Maximum/minimum (modulus) principle for holomorphic functions.
If
fis holomorphic on a domain\Omegaand attains a maximum on the boundary of\Omega, thenfis constant on\Omega.Except at
z_0\in \Omegawheref'(z_0)=0, iffattains a minimum on the boundary of\Omega, thenfis constant on\Omega.
Dirichlet problem for domain D
Let h: \partial D\to \mathbb{R} be a continuous function. Is there a harmonic function u on D such that u is continuous on \overline{D} and u|_{\partial D}=h?
We can always solve the problem for the unit disk.
u(z)=\frac{1}{2\pi}\int_0^{2\pi}h(e^{i t})\text{Re}\left(\frac{e^{it}+z}{e^{it}-z}\right)dt
Let z=re^{i\theta}
\text{Re}\left(\frac{e^{it}+re^{i\theta}}{e^{it}-re^{i\theta}}\right)=\frac {1-r^2}{1-2r\cos(\theta-t)+r^2}
This is called Poisson kernel.
Pr(\theta, t)>0 and \int_0^{2\pi}Pr(\theta, t)dt=1, \forall r,t
Chapter 8 Laurent series
when \sum_{n=-\infty}^{\infty}a_n(z-z_0)^n converges?
Claim \exists R>0 such that \sum_{n=-\infty}^{\infty}a_n(z-z_0)^n converges if |z-z_0|<R and diverges if |z-z_0|>R
Proof:
Let u=\frac{1}{z-z_0}
\sum_{n=0}^{\infty}a_n(z-z_0)^n has radius of convergence \frac{1}{R}
So the series converges if |u|<\frac{1}{R}
So |z-z_0|=\frac{1}{|u|}>\frac{1}{\frac{1}{R}}=R
QED
Laurent series
A Laurent series is a series of the form \sum_{n=-\infty}^{\infty}a_n(z-z_0)^n
The series converges in some annulus shape A=\{z:r_1<|z-z_0|<r_2\}
The annulus is called the region of convergence of the Laurent series.