2020-2ํ•™๊ธฐ, ๋Œ€ํ•™์—์„œ โ€˜์‘์šฉ๋ณต์†Œํ•จ์ˆ˜๋ก โ€™ ์ˆ˜์—…์„ ๋“ฃ๊ณ  ๊ณต๋ถ€ํ•œ ๋ฐ”๋ฅผ ์ •๋ฆฌํ•œ ๊ธ€์ž…๋‹ˆ๋‹ค. ์ง€์ ์€ ์–ธ์ œ๋‚˜ ํ™˜์˜์ž…๋‹ˆ๋‹ค :)

11 minute read

2020-2ํ•™๊ธฐ, ๋Œ€ํ•™์—์„œ โ€˜์‘์šฉ๋ณต์†Œํ•จ์ˆ˜๋ก โ€™ ์ˆ˜์—…์„ ๋“ฃ๊ณ  ๊ณต๋ถ€ํ•œ ๋ฐ”๋ฅผ ์ •๋ฆฌํ•œ ๊ธ€์ž…๋‹ˆ๋‹ค. ์ง€์ ์€ ์–ธ์ œ๋‚˜ ํ™˜์˜์ž…๋‹ˆ๋‹ค :)

Theorem. [Review] Cauchyโ€™s Integral Formula

Let $D$ be a simply connected domain, and $f$ be an analytic function in $D$.

Let $z_0 \in D$, and $C$ be any simple closed contour in $D$ that encloses $z_0$.

Then

\[\oint \frac{f(z)}{z-z_0} dz = 2\pi i f(z_0) \iff f(z_0) = \frac{1}{2\pi i} \oint \frac{f(z)}{z-z_0} dz\]

Extension of Cauchyโ€™s Integral Formula

Theorem.

Let $f(z)$ be analytic in $D$.

Then it has derivatives of all orders in $D$, which are then also analytic in $D$.

The derivatives are given by

\[f^{(n)}(z_0) = \frac{n!}{2\pi i} \oint_{C} \frac{f(z)}{(z-z_0)^{n+1}} dz\]

where $C$ is any simple closed CCW contour in $D$ encloseing $z_0$ such that its interior also lies in $D$.

Motivation.

analytic function $f(z)$๊ฐ€ ์•„๋ž˜์™€ ๊ฐ™์ด ์ฃผ์–ด์ ธ ์žˆ๋‹ค๊ณ  ํ•˜์ž.

\[f(z) = \frac{1}{2\pi i} \oint \frac{f(w)}{w-z} dz\]

์ด๋•Œ, $f(z)$๋ฅผ ๋ฏธ๋ถ„ํ•œ $\dfrac{d}{dz}f(z)$๋ฅผ ๊ตฌํ•  ๋•Œ, ์•„๋ž˜์™€ ๊ฐ™์ด ์ ๋ถ„ ๋‚ด๋ถ€์— ๋ฏธ๋ถ„ ์—ฐ์‚ฐ์ž $\dfrac{d}{dz}$๋ฅผ ๋ผ์›Œ๋„ฃ์„ ์ˆ˜ ์žˆ์„๊นŒ??

\[\frac{d}{dz} f(z) \overset{?}{=} \frac{1}{2\pi i} \oint \frac{d}{dz} \frac{f(w)}{w-z} dw\]

์ผ๋ฐ˜์ ์œผ๋กœ ์ ๋ถ„ ์—ฐ์‚ฐ์ž์™€ ๋ฏธ๋ถ„ ์—ฐ์‚ฐ์ž์˜ ๊ตํ™˜์€ ๋ถˆ๊ฐ€๋Šฅํ•˜๋‹ค! ํ•˜์ง€๋งŒ, ๋งŒ์•ฝ ์ด๊ฒƒ์ด ๊ฐ€๋Šฅํ•˜๋‹ค๋ฉด, ์šฐ๋ฆฌ๋Š” ์•„๋ž˜์™€ ๊ฐ™์€ ๊ฒฐ๊ณผ๋ฅผ ์–ป๋Š”๋‹ค.

\[\begin{aligned} \frac{d}{dz} f(z) &\overset{?}{=} \frac{1}{2\pi i} \oint \frac{d}{dz} \frac{f(w)}{w-z} dw \\ &= \frac{1}{2\pi i} \oint \frac{f(w)}{(w-z)^2} dw \end{aligned}\]

๋งŒ์•ฝ $n$๋ฒˆ ๋ฏธ๋ถ„ ํ•œ๋‹ค๋ฉด,

\[\left(\frac{d}{dz}\right)^n \frac{1}{w-z} = n! \frac{1}{(w-z)^{n+1}}\]

์ด๊ธฐ ๋•Œ๋ฌธ์— ์•„๋ž˜์™€ ๊ฐ™์€ ๊ฒฐ๊ณผ๋ฅผ ์–ป์„ ๊ฒƒ์ด๋‹ค.

\[\begin{aligned} f^{(n)}(z) &\overset{?}{=} \frac{1}{2\pi i} \oint \left(\frac{d}{dz}\right)^n \frac{f(w)}{w-z} dw \\ &= \frac{1}{2\pi i} \oint n! \frac{f(w)}{(w-z)^{n+1}} dw \\ &= \frac{n!}{2\pi i} \oint \frac{f(w)}{(w-z)^{n+1}} dw \end{aligned}\]

๋‹ค์‹œ ํ•œ๋ฒˆ ๋งํ•˜์ง€๋งŒ, ์ผ๋ฐ˜์ ์œผ๋กœ ์ ๋ถ„ ์—ฐ์‚ฐ์ž์™€ ๋ฏธ๋ถ„ ์—ฐ์‚ฐ์ž๋Š” ๊ตํ™˜๋˜์ง€ ์•Š๊ธฐ ๋•Œ๋ฌธ์— ์œ„์™€ ๊ฐ™์ด ์ฆ๋ช…ํ•˜๋ฉด ์•ˆ ๋œ๋‹ค! ์•„๋ž˜์— ๊ทนํ•œ์˜ ๊ฐœ๋…์„ ์ด์šฉํ•ด ์ด๊ฒƒ์„ ์ œ๋Œ€๋กœ ์ฆ๋ช…ํ•œ๋‹ค.


Proof.

<Extended Cauchy Integral>๋ฅผ ์ฆ๋ช…ํ•˜๊ธฐ ์œ„ํ•ด $f(z)$๋ฅผ ํ•œ๋ฒˆ ๋ฏธ๋ถ„ํ•œ ๊ฒฐ๊ณผ๊ฐ€ ์•„๋ž˜๊ฐ€ ๋จ์„ ๋ณด์ด์ž. ๋‘๋ฒˆ, ์„ธ๋ฒˆ ๋ฏธ๋ถ„์— ๋Œ€ํ•œ ๊ฒฐ๊ณผ๋Š” $fโ€™(z)$๋ฅผ ์ฆ๋ช…ํ•œ ๋ฐฉ์‹์„ ๊ทธ๋Œ€๋กœ ์‚ฌ์šฉํ•˜๋ฉด ๋œ๋‹ค.

Let assume $C$ a circle with radius $r$.

\[\begin{equation} f'(z_0) = \frac{1}{2\pi i} \oint_{C} \frac{f(z)}{(z-z_0)^2} dz \end{equation}\]

๋จผ์ € ์ •์˜์— ๋”ฐ๋ผ $fโ€™(z_0)$๋Š” ์•„๋ž˜์™€ ๊ฐ™๋‹ค.

\[f'(z) = \lim_{h \rightarrow 0} \frac{f(z_0 + h) - f(z_0)}{h}\]

์ด๋•Œ, ์‹ (1)์—์„œ ์šฐ๋ณ€์ด ์ ๋ถ„์ด๋‹ˆ ์ขŒ๋ณ€์ธ $fโ€™(z)$๋„ ์ ๋ถ„์˜ ํ˜•ํƒœ๋กœ ๋ฐ”๊ฟ”์ฃผ์ž. ์ด ๊ณผ์ •์—์„œ <Cauchy Integral>์„ ์‚ฌ์šฉํ•œ๋‹ค.

\[\begin{equation} f'(z) = \frac{1}{h} \left( \frac{1}{2\pi i} \oint_{C} \frac{f(z)}{z-(z_0 + h)} dz - \frac{1}{2\pi i} \oint_{C} \frac{f(z)}{z-z_0} dz \right) \end{equation}\]

(์ด๋•Œ, $h$๋Š” ์ถฉ๋ถ„ํžˆ ์ž‘์•„์„œ $\left| h \right| < r/2$์ด๋‹ค.)

์‹ (2)์˜ ์ ๋ถ„์‹์„ ์กฐ๊ธˆ ๋‹ค๋“ฌ์–ด๋ณด์ž.

\[\begin{equation} \begin{aligned} (2) &= \frac{1}{h} \frac{1}{2\pi i} \oint_C f(z) \left(\frac{1}{z-(z_0+h)} - \frac{1}{z-z_0} \right) dz \\ &= \frac{1}{h} \frac{1}{2\pi i} \oint_C f(z) \left( \frac{(z-z_0) - (z-(z_0+h)))}{\left(z-(z_0+h)\right)(z-z_0)} \right) dz \\ &= \frac{1}{\cancel{h}} \frac{1}{2\pi i} \oint_C f(z) \frac{\cancel{h}}{\left(z-(z_0+h)\right)(z-z_0)} dz \\ &= \frac{1}{2\pi i} \oint_C \frac{f(z)}{\left(z-(z_0+h)\right)(z-z_0)} dz \end{aligned} \end{equation}\]

์ด์ œ ์‹ (3)์„ ์‹ (1)์— ์ ์šฉํ•˜๊ณ , ๊ทนํ•œ์„ ์ทจํ•˜๊ธฐ ํŽธํ•œ ๊ผด๋กœ ๋ฐ”๊ฟ”์ฃผ์ž.

\[\begin{equation} \begin{aligned} &f'(z_0) - \frac{1}{2\pi i} \oint_{C} \frac{f(z)}{(z-z_0)^2} dz \\ &= \frac{1}{2\pi i} \oint_C \frac{f(z)}{\left(z-(z_0+h)\right)(z-z_0)} dz - \frac{1}{2\pi i} \oint_{C} \frac{f(z)}{(z-z_0)^2} dz \\ &= \frac{1}{2\pi i} \oint_C f(z) \left( \frac{1}{\left(z-(z_0+h)\right)(z-z_0)} - \frac{1}{(z-z_0)^2} \right) dz \\ &= \frac{1}{2\pi i} \oint_C f(z) \left( \frac{(z-z_0)^2 - ((z-z_0) - h)(z-z_0)}{\left(z-(z_0+h)\right)(z-z_0)(z-z_0)^2} \right) dz \\ &= \frac{1}{2\pi i} \oint_C f(z) \left( \frac{h\cancel{(z-z_0)}}{\left(z-(z_0+h)\right)\cancel{(z-z_0)}(z-z_0)^2} \right) dz \\ &= \frac{1}{2\pi i} \oint_C \frac{f(z) \cdot h}{\left(z-(z_0+h)\right)(z-z_0)^2} dz \end{aligned} \end{equation}\]

์ด์ œ ์‹ (4)๊ฐ€ 0์œผ๋กœ ์ˆ˜๋ ดํ•จ์„ ๋ณด์ด๋ฉด <Extended Cauchy Integral>์„ ์ฆ๋ช…ํ•˜๊ฒŒ ๋œ๋‹ค! ์ด๊ฒƒ์€ <ML-Inequality>๋ฅผ ์‚ฌ์šฉํ•˜๋ฉด ๋œ๋‹ค.

์ด๋•Œ, $\left| z-(z_0+h) \right| \ge r/2$์ด๊ณ , $\left| f(z) \right| \le M$ for some $M$ ์ด๋ฏ€๋กœ,

\[\begin{equation} \left| \frac{1}{2\pi i} \oint_C \frac{f(z) \cdot h}{\left(z-(z_0+h)\right)(z-z_0)^2} dz \right| \le \left( \frac{1}{2\pi} \frac{M \cdot h}{(r/2) \; r^2}\right) \cdot \left( 2\pi r \right) = 2 \; \frac{M \cdot h}{r^2} \end{equation}\]

$h \rightarrow 0$์ผ ๋•Œ, ์‹ (5)๊ฐ€ 0์œผ๋กœ ์ˆ˜๋ ดํ•˜๋ฏ€๋กœ ์‹ (1)์ด ์„ฑ๋ฆฝํ•œ๋‹ค! $\blacksquare$


Application of Cauchy Integral

Cauchyโ€™s Inequality

Theorem.

Let $f(z)$ be an analytic inside and on a positively oriented circle $C$ or radius $r$ and center $z_0$.

If $\left| f(z) \right| \le M$ on $C$, then

\[\left| f^{(n)}(z_0) \right| \le \frac{n!M}{R^n}\]


Proof.

์ฆ๋ช…์€ ๊ฐ„๋‹จํ•˜๋‹ค. <Extended Cauchy Integral>์— ๋ฐ”๋กœ <ML-Inequality>๋ฅผ ์ทจํ•ด์ฃผ๋ฉด ๋œ๋‹ค.

\[\left| f^{(n)}(z_0) \right| = \left| \frac{n!}{2\pi i} \oint_C \frac{f(z)}{(z-z_0)^n} dz \right| \le \left( \frac{n!}{2\pi} \frac{M}{R^{n+1}} \right) \left( 2\pi R \right) = \frac{n!M}{R^n}\]

$\blacksquare$

Liouvilleโ€™s Theorem

Theorem.

If a function $f(z)$ is entire and bounded, then $f(z)$ is constant.

<๋ฆฌ์šฐ๋นŒ์˜ ์ •๋ฆฌ>๋Š” ๋Œ€์šฐ๋„ ํ•จ๊ป˜ ์‚ดํŽด๋ณด๋ฉด ์ข‹๋‹ค.

(๋Œ€์šฐ) If $f(z)$ is not constant, then $f(z)$ is not bounded (=explode to $+\infty$ or $-\infty$).


Proof.

Supp. $f(z)$ is bounded, then $\left| f(z) \right| \le M$ for all $z$.

Consider a circle $C_R$ with radius $R$. Then, due to

\[\left| f'(z_0) \right| \le \frac{M}{R}\]

$\left| fโ€™(z_0) \right| \le 0$ as $R \rightarrow \infty$

Therefore, $fโ€™(z_0) = 0$ for all $z_0 \in \mathbb{C}$.

This means $f(z)$ is constant. $\blacksquare$

Fundamental Theorem of Algebra

Theorem.

Any polynomial

\[P(z) = a_n z^n + a_{n-1} z^{n-1} + \cdots + a_1 z + a_0, \quad (a_n \ne 0)\]

of degree $n$ ($n \ge 1$) has a solution.


Proof.

(๊ท€๋ฅ˜๋ฒ•) Supp. that $P(z)$ has no solution.

Then $\dfrac{1}{P(z)}$ is entire, because thereโ€™s no $z_0$ s.t. $p(z_0) = 0$

It is also bounded in the complex plane!

\[\begin{aligned} \left| P(z) \right| &= \left| a_n z^n + a_{n-1} z^{n-1} + \cdots + a_1 z + a_0 \right| \\ & \ge \left| a_n z^n \right| - \left| a_{n-1} z^{n-1} + \cdots + a_1 z + a_0 \right| \end{aligned}\]

์ด๋•Œ, $\left| a_{n-1} z^{n-1} + \cdots + a_1 z + a_0 \right|$์— ์‚ผ๊ฐ ๋ถ€๋“ฑ์‹์„ ์ ์šฉํ•˜๋ฉด ์•„๋ž˜์™€ ๊ฐ™๋‹ค.

\[\left| a_{n-1} z^{n-1} + \cdots + a_1 z + a_0 \right| \le \left| a_{n-1} \right| \left|z \right|^{n-1} + \cdots + \left| a_1 \right| \left| z \right| + \left| a_0 \right|\]

์ด๊ฒƒ์„ ์ ์šฉํ•˜๋ฉด,

\[\begin{aligned} \left| P(z) \right| &\ge \left| a_n z^n \right| - \left| a_{n-1} z^{n-1} + \cdots + a_1 z + a_0 \right| \\ & \ge \left| a_n \right| \left| z \right|^n - \left| a_{n-1} \right| \left|z \right|^{n-1} - \cdots - \left| a_1 \right| \left| z \right| - \left| a_0 \right| \end{aligned}\]

์ด๋•Œ, ๋งจ์•ž์— ์žˆ๋Š” $\left| a_n \right| \left| z \right|^n$์˜ power์— ์˜ํ•ด ์–ด๋–ค ์–‘์ˆ˜ $R$๊ฐ€ ์กด์žฌํ•ด $\left| z \right| \ge R$์—์„  $\left| P(z) \right| \ge 1$๊ฐ€ ์„ฑ๋ฆฝํ•œ๋‹ค.

๊ทธ๋ฆฌ๊ณ  ์ด๊ฒƒ์€ $\left| \dfrac{1}{P(z)} \right|$์„ bounded ์‹œํ‚จ๋‹ค.

\[\left| P(z) \right| \ge 1 \iff \left| \frac{1}{P(z)} \right| \le 1\]

๋ฐ˜๋Œ€๋กœ $\left| z \right| < R$์—์„  $\dfrac{1}{P(z)}$๊ฐ€ continuousํ•˜๊ณ , $P(z) = 0$์ด ๋˜๋Š” ์ง€์ ์ด ์กด์žฌํ•˜์ง€ ์•Š๊ธฐ ๋•Œ๋ฌธ์— $\left| \dfrac{1}{P(z)} \right|$๋„ bounded ๋˜์–ด ์žˆ๋‹ค!

๋”ฐ๋ผ์„œ, $\dfrac{1}{P(z)}$๋Š” bounded ๋œ entire function์ด๋‹ค. ๊ทธ๋ ‡๋‹ค๋ฉด, ์•ž์—์„œ๋ณด์ธ <๋ฆฌ์šฐ๋นŒ์˜ ์ •๋ฆฌ>์— ์˜ํ•ด $\dfrac{1}{P(z)}$๋Š” constant function์ด ๋œ๋‹ค. ํ•˜์ง€๋งŒ, $n \ge 0$์ด๋ฏ€๋กœ $\dfrac{1}{P(z)}$๋Š” ๋ช…๋ฐฑํžˆ constant function์ด ์•„๋‹ˆ๋‹ค!!

๋”ฐ๋ผ์„œ, โ€œ$P(z)$ has no solution.โ€๋ผ๋Š” ์ฒ˜์Œ์˜ ๊ฐ€์ •์€ ๊ฑฐ์ง“์ด๋‹ค. $P(z)$๋Š” ์ ์–ด๋„ ํ•˜๋‚˜์˜ solution์„ ๊ฐ€์ง„๋‹ค!

๋งŒ์•ฝ $P(z)$๊ฐ€ ์ ์–ด๋„ ํ•˜๋‚˜์˜ solution์„ ๊ฐ€์ง„๋‹ค๋ฉด, $P(z)$๋ฅผ ์•„๋ž˜์™€ ๊ฐ™์ด ๊ธฐ์ˆ ํ•  ์ˆ˜ ์žˆ๋‹ค.

\[P(z) = (z-z_1) Q_1(z)\]

์ด๋•Œ, $Q_1(z)$๋Š” $n-1$ ์ฐจ์ˆ˜๋ฅผ ๊ฐ€์ง€๋Š” polynomial์ด๋‹ค. $Q_1(z)$์— ๋Œ€ํ•ด ์•ž์„  ๋…ผ์˜๋ฅผ ๋‹ค์‹œ ์ ์šฉํ•˜๋ฉด, $Q_1(z)$์ด ์ ์–ด๋„ ํ•˜๋‚˜์˜ solution์„ ๊ฐ€์ง์„ ์•Œ ์ˆ˜ ์žˆ๋‹ค. ๋”ฐ๋ผ์„œ

\[P(z) = (z-z_1)(z-z_2) Q_2(z)\]

์ด๊ฒƒ์„ ๋ฐ˜๋ณตํ•˜๋ฉด, $P(z)$๊ฐ€ $n$๊ฐœ solution์„ ๊ฐ€์ง์„ ํ™•์ธํ•  ์ˆ˜ ์žˆ๋‹ค!! $\blacksquare$

Moreraโ€™s Theorem

Theorem.

Let $f(z)$ be a continuous function in a domain $D$.

If

\[\oint_C f(z) dz = 0\]

for any closed contour in $D$, then $f(z)$ is analytic in $D$.

<๋ชจ๋ ˆ๋ผ์˜ ์ •๋ฆฌ>๋Š” <Cauchy-Goursat Theorem>์˜ ์—ญ์„ ๊ธฐ์ˆ ํ•œ ์ •๋ฆฌ๋‹ค.


Proof.

\[\oint_C f(z) dz = 0 \implies f(z) \; \text{is primitive.}\]

๋”ฐ๋ผ์„œ, $f(z)$์— ๋Œ€ํ•œ Anti-Derivative Function $F(z)$๊ฐ€ ์กด์žฌํ•ด ์•„๋ž˜๊ฐ€ ์„ฑ๋ฆฝํ•œ๋‹ค.

\[F'(z) = f(z)\]

์ด๋•Œ, $F(z)$์˜ derivative์ธ $f(z)$๊ฐ€ continuous ํ•˜๋ฏ€๋กœ $F(z)$๋Š” analytic function์ด๋‹ค. Anti-Derivative $F(z)$๊ฐ€ analytic์ด๋ฏ€๋กœ derivative์ธ $f(z)$ ์—ญ์‹œ analytic์ด๋‹ค! $\blacksquare$