Files
NoteNextra-origin/pages/Math416/Math416_L23.md
2025-04-17 12:13:46 -05:00

161 lines
4.5 KiB
Markdown

# Math416 Lecture 23
## Chapter 9: Generalized Cauchy Theorem
### Separation lemma
Let $\Omega$ be an open subset in $\mathbb{C}$, let $K\subset \Omega$ be compact. Then There exists a simple contour $\Gamma$ such that
$$
K\subset \text{int}(\Gamma)\subset \Omega
$$
#### Corollary 9.9 for separation lemma
Let $\Gamma$ be the contour constructed in the separation lemma. Let $f\in O(\Omega)$ be holomorphic on $\Omega$. Then $\forall z_0\in K$ such that
$$
f(z_0)=\frac{1}{2\pi i}\int_{\Gamma}\frac{f(z)}{z-z_0}dz
$$
Proof:
Suppose $h\in O(G)$, then $\int_{\partial S} h(z)dz=0$, by Cauchy's theorem for square, followed from the triangle case.
So $\int_{\Gamma} h(z)dz=0=\sum_{j=1}^n \int_{\partial S_j} h(z)dz$
Fix $z_0\in K$,
$$
g(z_0)=\begin{cases}
\frac{f(z)-f(z_0)}{z-z_0} & z\neq z_0 \\
f'(z_0) & z=z_0
\end{cases}
$$
So $\int_{\Gamma} g(z)dz=0$
Thus
$$
\begin{aligned}
\int_{\Gamma}\frac{f(z)}{z-z_0}dz-\int_{\Gamma}\frac{f(z_0)}{z-z_0}dz&=0 \\
\int_{\Gamma}\frac{f(z)}{z-z_0}dz&=f(z_0)\int_{\Gamma}\frac{1}{z-z_0}dz \\
&=f(z_0)\cdot 2\pi i
\end{aligned}
$$
QED
#### Theorem 9.10 Cauchy's Theorem
Let $\Omega$ be an open subset in $\mathbb{C}$, let $\Gamma$ be a contour with $int(\Gamma)\subset \Omega$. Let $f\in O(\Omega)$ be holomorphic on $\Omega$. Then
$$
\int_{\Gamma} f(z)dz=0
$$
Proof:
Let $K\subset \mathbb{C}\setminus \text{ext}(\Gamma)$.
By separation lemma, $\exists \Gamma_1$ s.t. $K\subset \text{int}(\Gamma_1)\subset \Omega$.
Notice that Separation lemma ensured that $w\neq z$ for all $w\in \Gamma_1, z\in \Gamma$.
By Corollary 9.9, $\forall z\in K, f(z)=\frac{1}{2\pi i}\int_{\Gamma_1}\frac{f(w)}{w-z}dw$
$$
\int_{\Gamma} f(z)dz=\frac{1}{2\pi i}\int_{\Gamma}\left[\int_{\Gamma_1}\frac{f(w)}{w-z}dw\right]dz
$$
By Fubini's theorem (In graduate course for analysis),
$$
\begin{aligned}
\int_{\Gamma} f(z)dz&=\frac{1}{2\pi i}\int_{\Gamma_1}\left[\int_{\Gamma}\frac{f(w)}{w-z}dz\right]dw \\
&=\frac{1}{2\pi i}\int_{\Gamma_1}f(w)\left[\int_{\Gamma}\frac{1}{w-z}dz\right]dw \\
&=\frac{1}{2\pi i}\int_{\Gamma_1}f(w)\cdot 2\pi i \ \text{ind}_{\Gamma}(w)dw \\
&=0
\end{aligned}
$$
Since the winding number for $\Gamma$ on $w\in \Gamma_1$ is 0. ($w$ is outside of $\Gamma$)
QED
### Homotopy
Suppose $\gamma_0, \gamma_1$ are two curves from
$[0,1]$ to $\Omega$ with same end points $P,Q$.
A homotopy is a continuous function of curves $\gamma_t, 0\leq t\leq 1$, deforming $\gamma_0$ to $\gamma_1$, keeping the end points fixed.
Formally, if $H:[0,1]\times [0,1]\to \Omega$ is a continuous function satsifying
1. $H(s,0)=\gamma_0(s)$, $\forall s\in [0,1]$
2. $H(s,1)=\gamma_1(s)$, $\forall s\in [0,1]$
3. $H(0,t)=P$, $\forall t\in [0,1]$
4. $H(1,t)=Q$, $\forall t\in [0,1]$
Then we say $H$ is a homotopy between $\gamma_0$ and $\gamma_1$. (If $\gamma_0$ and $\gamma_1$ are closed curves, $Q=P$)
#### Lemma 9.12 Technical Lemma
Let $\phi:[0,1]\times [0,1]\to \mathbb{C}\setminus \{0\}$ is continuous. Then there exists a continuous map $\psi:[0,1]\times [0,1]\to \mathbb{C}$ such that $e^\phi=\psi$. Moreover, $\psi$ is unique up to an additive constant in $2\pi i\mathbb{Z}$.
Proof:
Let $\phi_t(s)=\phi(s,t)$, $0\leq t\leq 1$.
Then $\exists \psi_{00}$ such that $e^{\psi_{00}(s)}=\phi(0,t)$.
$\exists \psi_{t}(s)$ such that $e^{\psi_{t}(s)}=\phi_t(s)$.
We want to show $\psi_t(s)$ is continuous in $t$.
Since $\exists \epsilon>0$ such that $\phi(s,t)$ is at least $\epsilon$ away from $0$ for all $s\in [0,1]$ and $t\in [0,1]$.
Moreover, $\phi(s,t)$ is uniformly continuous.
So $\exists \delta>0$ such that $|\phi(s,t)-\phi(s,t_0)|<\epsilon$ if $|t-t_0|<\delta$.
Therefore,
$$
\begin{aligned}
\left|\frac{\phi(s,t)}{\phi(s,t_0)}-1\right|&<\frac{\epsilon}{\phi(s,t_0)}
&<1
\end{aligned}
$$
So $\text {Re} \frac{\phi(s,t)}{\phi(s,t_0)}>0$.
Therefore, $\text{Log} \frac{\phi(s,t)}{\phi(s,t_0)}=\chi(s,t)$ is continuous on $s\in [0,1], t\in [t_0-\delta, t_0+\delta]$.
So $e^{\chi(s,t)}=\frac{\phi(s,t)}{\phi(s,t_0)}$, $\chi(s,t_0)=0,\forall s\in [0,1]$
Define $\tilde{\psi}(s,t)=\chi(s,t)+\chi(s,t_0)$. So this function is continuous.
And $e^{\tilde{\psi}(s,t)}=e^{\chi(s,t)+\chi(s,t_0)}=e^{\chi(s,t)}\cdot e^{\chi(s,t_0)}=\phi(s,t)$.
$$
\begin{aligned}
\tilde{\psi}(0,t_0)&=\chi(0,t_0)+\psi(0,t_0) \\
&=0+\psi_{00}(t_0) \\
&=\psi_{00}(t_0)
\end{aligned}
$$
$\tilde{\psi}(s,0)$ and $\psi(t,0)$ on $t\in[t_0-\delta, t_0+\delta]$ are both logs of the same function, and agree to each other on $t_0$.
Therefore, $\tilde{\psi}(s,0)=\psi(s,0)+\text{const}$
QED
#### Theorem 9.13 Cauchy's Theorem for Homotopic Curves