2020-2ํ•™๊ธฐ, ๋Œ€ํ•™์—์„œ โ€˜ํ˜„๋Œ€๋Œ€์ˆ˜1โ€™ ์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€˜ํ˜„๋Œ€๋Œ€์ˆ˜1โ€™ ์ˆ˜์—…์„ ๋“ฃ๊ณ  ๊ณต๋ถ€ํ•œ ๋ฐ”๋ฅผ ์ •๋ฆฌํ•œ ๊ธ€์ž…๋‹ˆ๋‹ค. ์ง€์ ์€ ์–ธ์ œ๋‚˜ ํ™˜์˜์ž…๋‹ˆ๋‹ค :)


Gaussian Integer๋Š” โ€œEuclidean Domainโ€์˜ ์ผ์ข…์ด๋‹ค. Euclidean Domain์— ๋Œ€ํ•œ ํฌ์ŠคํŠธ๋Š” ์ด๊ณณ์—์„œ ํ™•์ธํ•  ์ˆ˜ ์žˆ๋‹ค.

  1. Gaussian Integers
  2. Multiplicative Norms



Gaussian Integers


Definition.

A Gaussian Integer is a complex number $a + bi$, where $a, b \in \mathbb{Z}$.

For Gaussian integer $\alpha = a + bi$, the norm $N(\alpha) = a^2 + b^2$.

์œ„์˜ Gaussian Integer๋ฅผ ๋ชจ๋‘ ๋ชจ์€ ์ง‘ํ•ฉ์ด ๋ฐ”๋กœ $\mathbb{Z}[i] \subset \mathbb{C}$๊ฐ€ ๋œ๋‹ค.

์šฐ๋ฆฌ์˜ ๋ชฉํ‘œ๋Š” โ€œGaussian Integers $\mathbb{Z}[i]$๊ฐ€ Euclidean Domain์ด ๋จโ€์„ ๋ณด์ด๋Š” ๊ฒƒ์ด๋‹ค!




Lemma 47.2

On $\mathbb{Z}[i]$, the following properties of Norm holds.

  1. $N(\alpha) \ge 0$
  2. $N(\alpha) = 0 \iff \alpha = 0$
  3. $N(\alpha \beta) = N(\alpha)N(\beta)$

์ฆ‰, $N$์€ semi-group homomoprhism์ด๋‹ค.

Gaussian Norm $N$์„ ์ž˜ ์ƒ๊ฐํ•ด๋ณด๋ฉด, ๋„ˆ๋ฌด ๋‹น์—ฐํ•œ ๋ช…์ œ๋“ค์ด๋‹ค.


Lemma 47.3

$Z[i]$ is an Integral Domain.

proof.

$\mathbb{Z}[i]$๋Š” commutative ring with unity์ด๋‹ค.

$\mathbb{Z}[i]$๊ฐ€ Integral Domain์ž„์„ ๋ณด์ด๊ธฐ ์œ„ํ•ด zero-divisor๊ฐ€ ์กด์žฌํ•˜์ง€ ์•Š์Œ์„ ๋ณด์—ฌ์•ผ ํ•œ๋‹ค.

โ€œLemma 47.2โ€์— ์˜ํ•ด

if $\alpha \beta = 0$, then

\[N(\alpha)N(\beta) = N(\alpha \beta) = N(0) = 0\]

๋”ฐ๋ผ์„œ $\alpha \beta = 0$์€ $N(\alpha) = 0$ ๋˜๋Š” $N(\beta) = 0$์„ ์˜๋ฏธํ•˜ํ•œ๋‹ค.

๋‹ค์‹œ โ€œLemma 47.2โ€์— ์˜ํ•ด ์œ„์˜ ๊ฒฐ๊ณผ๋Š” $\alpha = 0$ ๋˜๋Š” $\beta = 0$์„ ์˜๋ฏธํ•œ๋‹ค.

์ฆ‰, zero-divisor๊ฐ€ ์กด์žฌํ•˜์ง€ ์•Š์œผ๋ฏ€๋กœ $\mathbb{Z}[i]$๋Š” Integral Domain์ด๋‹ค.


Theorem 47.4

The function $\nu$ given by $\nu(\alpha) = N(\alpha)$ for non-zero $\alpha \in \mathbb{Z}[i]$ is an Euclidean norm on $\mathbb{Z}[i]$.

Thus $\mathbb{Z}[i]$ is an Euclidean Domain.

proof.

Note that for $\beta = b_1 + b_2 i \ne 0$, $N(\beta) = {b_1}^2 + {b_2}^2$, so $N(\beta) \ge 1$.

Then for all non-zero $\alpha, \beta \in \mathbb{Z}[i]$, $N(\alpha) \le N(\alpha)N(\beta) = N(\alpha \beta)$.

This proves Condition 2 for a Euclidean norm.


์ด์ œ Euclidean norm์˜ ์ฒซ๋ฒˆ์จฐ ์กฐ๊ฑด์ธ โ€œdivision algorithmโ€์— ๋Œ€ํ•ด ์ฆ๋ช…ํ•ด์•ผ ํ•œ๋‹ค.

Let $\alpha, \beta \in \mathbb{Z}[i]$ with $\alpha = a_1 + a_2 i$, $\beta = b_1 + b_2 i \ne 0$.

We must find $\sigma, \rho \in \mathbb{Z}[i]$ s.t. $\alpha = \beta \sigma \rho$, where either $\rho = 0$ or $N(\rho) < N(\beta)$.

Let $\alpha / \beta = r + si$ for $r, s \in \mathbb{Q}$. (by $\mathbb{C}$ ์•„๋ž˜ ์—ฐ์‚ฐ)


Let $n, m \in \mathbb{Z}$ as close as possible to the rational numbers $r$ and $s$.

Let $\sigma = n + m i$ and $\rho = \alpha - \beta \sigma$.

If $\rho = 0$, we are done.

Otherwise, by construction of $\sigma$, we see that $\left| r - n \right| \le \frac{1}{2}$ and $\left| s - m \right| \le \frac{1}{2}$

Therefore,

\[\begin{aligned} N \left( \frac{\alpha}{\beta} - \sigma \right) &= N \left( (r+si) - (n+mi) \right) \\ &= N \left( (r - n) - (s - m)i \right) \le \left( \frac{1}{2} \right)^2 + \left( \frac{1}{2} \right)^2 = \frac{1}{2} \end{aligned}\]

Thus, we obtain

\[\begin{aligned} N(\rho) &= N(\alpha - \beta \sigma) = N\left( \beta \left( \frac{\alpha}{\beta} - \sigma \right) \right) \\ &= N(\beta)N\left( \frac{\alpha}{\beta} - \sigma \right) \le N(\beta) \frac{1}{2} < N(\beta) \end{aligned}\]

so, $N(\rho) < N(\beta)$.

Therefore, Gaussian norm $N$ is an Euclidean norm. $\blacksquare$


โ€$\mathcal{U}(\mathbb{Z}[i]) = \{ \pm 1, \pm i\}$โ€

๋‹น์—ฐํžˆ ๊ทธ๋ ‡๊ฒ ์ง€๋งŒ, $\mathbb{Z}[i]$๋Š” $\mathbb{Z}$์™€๋Š” ๋‹ค๋ฅธ ๋ชจ์Šต์ด ๋ฐœ๊ฒฌ๋œ๋‹ค.

์˜ˆ๋ฅผ ๋“ค์–ด $\mathbb{Z}$์—์„  5๊ฐ€ irreducible์ธ ๋ฐ˜๋ฉด,

$\mathbb{Z}[i]$์—์„  5๊ฐ€ $5 = (1 + 2i)(1 - 2i)$๋กœ ๋ถ„ํ•ด๊ฐ€๋Šฅํ•˜๋‹ค!



Multiplicative Norms

์ด๋ฒˆ ์„น์…˜์—์„œ๋Š” ์„ธ์‹ฌํ•˜๊ฒŒ ์ •์˜๋œ norm์€ Integral Domain $D$์˜ arithmetic structure๋ฅผ ๊ฒฐ์ •ํ•˜๋Š” ๋ฐ์— ๋งŽ์€ ๋„์›€์„ ์ค€๋‹ค๋Š” ์‚ฌ์‹ค์„ ์‚ดํŽด๋ณผ ๊ฒƒ์ด๋‹ค.

Algebraic Number Theory์—์„  ์ด๋ ‡๊ฒŒ norm ์„ ํ†ตํ•ด ๋Œ€์ˆ˜์ ์ธ ๊ตฌ์กฐ๋ฅผ ํŒŒ์•…ํ•˜๋Š” ๊ฒƒ์ด ๋นˆ๋ฒˆํ•˜๋‹ค.


Definition.

Let $D$ be an integral domain.

A multiplicative norm $N$ on $D$ is a function mapping $D$ into the integers $\mathbb{Z}$ such that the following conditions are satisfied:

  1. $N(\alpha) = 0 \iff \alpha = 0$
  2. $N(\alpha \beta) = N(\alpha) N(\beta)$ for all $\alpha, \beta \in D$


Theorem 47.7

If $D$ is an integral domain with a multiplicative norm $N$,

then

  1. $N(1) = 1$
  2. $\left| N(u) \right| = 1$ for every unit $u \in D$.
  3. If Every $\alpha$ s.t. $\left| N(\alpha) \right| = 1$ is a unit in $D$,
    then an elt $\pi$ in $D$ with $\left| N(\pi) \right| = p$ for a prime $p$ is an irreducible of $D$. ๐Ÿ”ฅ

proof.

Let $D$ be an integral domain with a multiplicative norm $N$.

(1๋ฒˆ ๋ช…์ œ)

\[N(1) = N(1 \cdot 1) = N(1) N(1)\]

show that $N(1) = 1$


(2๋ฒˆ ๋ช…์ œ)

If $u$ is a unit in $D$, then

\[1 = N(1) = N(u u^{-1}) = N(u) N(u^{-1})\]

Since $N(u)$ is an integer, this implies that $\left| N(u) \right| = 1$.


๐Ÿ”ฅ 3๋ฒˆ ๋ช…์ œ ๐Ÿ”ฅ

Supp. that the units of $D$ are exactly the elements of norm $\pm 1$.

For $\pi \in D$ with $\left| N(\pi) \right| = p$ where $p$ is a prime.

Then if $\pi = \alpha \beta$, we have

\[p = \left| N(\pi) \right| = \left| N(\alpha) N(\beta) \right|\]

so either $\left| N(\alpha) \right| = 1$ or $\left| N(\beta) \right| = 1$.

By our assumption, this means $\alpha$ or $\beta$ is a unit.

Thus $\pi$ is an irreducible of $D$.


Examples.

In Gaussian Integers $\mathbb{Z}[i]$, $N(\alpha)$ is a multiplicative norm!

$1+2i$ and $1-2i$ are irreducibles.

๋ฐ˜๋Œ€๋กœ $5$์˜ ๊ฒฝ์šฐ $N(5) = 25$์ด๊ธฐ ๋•Œ๋ฌธ์— reducible์ด ๋œ๋‹ค.




Example. Integral Domain, but not UFD ๐Ÿ”ฅ

Let $\mathbb{Z}[\sqrt{-5}] = \{ a+ib\sqrt{5} \mid a, b \in \mathbb{Z} \}$.

As a subset of the complex numbers $\mathbb{Z}[\sqrt{-5}]$ is an integral domain.

Define $N$ on $\mathbb{Z}[\sqrt{-5}]$ by

\[N(a + b\sqrt{-5}) = a^2 + 5 b^2\]

Clearly, $N(\alpha) = 0$ $\iff$ $\alpha = 0$.

Also $N(\alpha \beta) = N(\alpha) N(\beta)$.

์ด๋ฒˆ์—๋Š” multiplicative norm์˜ ํŠน์ง•์— ๋น„์ถ”์–ด $\mathbb{Z}[\sqrt{-5}]$์˜ unit๋ฅผ ์ƒ๊ฐํ•ด๋ณด์ž.

$N(\alpha) = a^2 + 5 b^2 = 1$์ด ๋˜๋Š” $\alpha$๋Š” ์˜ค์ง $b=0$์ด๊ณ  $a = \pm 1$์ด์–ด์•ผ ํ•œ๋‹ค.

๋”ฐ๋ผ์„œ $\mathcal{U}(\mathbb{Z}[\sqrt{-5}]) = \{ \pm 1\}$์ด๋‹ค.

ย  ์ด์ œ $\mathbb{Z}[\sqrt{-5}]$์˜ ์›์†Œ์ธ 21์— ๋Œ€ํ•œ factorization์„ ์‚ดํŽด๋ณด์ž.

21์€ ๋‘ ๊ฐ€์ง€ factorization์„ ๊ฐ€์งˆ ์ˆ˜ ์žˆ๋‹ค.

  • $21 = (3)(7)$
  • $21 = (1+2\sqrt{-5})(1-2\sqrt{-5})$

์ด์ œ $3$, $7$, $1+2\sqrt{-5}$, $1-2\sqrt{-5}$๊ฐ€ irreducible์ž„๋งŒ ๋ณด์ธ๋‹ค๋ฉด, $\mathbb{Z}[\sqrt{-5}]$๊ฐ€ UFD๊ฐ€ ์•„๋‹˜์„ ๋ณด์ผ ์ˆ˜ ์žˆ๋‹ค.


($3$์˜ ๊ฒฝ์šฐ)

Supp. that $3 = \alpha \beta$, then

\[9 = N(3) = N(\alpha) N(\beta)\]

$N(\alpha)$๊ฐ€ ๊ฐ€์งˆ ์ˆ˜ ์žˆ๋Š” ๊ฐ’์€ 1 ๋˜๋Š” 3 ๋˜๋Š” 9์ด๋‹ค.

  • If $N(\alpha) = 1$, then $\alpha$ is a unit.
  • $N(\alpha) = a^2 + 5 b^2$์ด๋ฏ€๋กœ $N(\alpha) = 3$๋ฅผ ๋งŒ์กฑ์‹œํ‚ฌ ์ˆ˜ ์žˆ๋Š” $\alpha$๋Š” ์กด์žฌํ•˜์ง€ ์•Š๋Š”๋‹ค.
  • If $N(\alpha) = 9$, $\beta$ is a unit.

๋”ฐ๋ผ์„œ $3$์€ irreducible์ด๋‹ค.

๋น„์Šทํ•œ ๋ฐฉ๋ฒ•์œผ๋กœ $7$๊ฐ€ irreducible์ž„๋„ ๋ณด์ผ ์ˆ˜ ์žˆ๋‹ค.


($1 + 2\sqrt{-5}$์˜ ๊ฒฝ์šฐ)

If $1 + 2\sqrt{-5} = \gamma \delta$, we have

\[21 = N(1 + 2\sqrt{-5}) = N(\gamma)N(\delta)\]

๋”ฐ๋ผ์„œ $N(\gamma)$๋Š” 1, 3, 7, ๋˜๋Š” 21์˜ ๊ฐ’์„ ๊ฐ–๋Š”๋‹ค.

์•ž์—์„œ 3๊ณผ 7์„ norm์œผ๋กœ ๊ฐ–๋Š” ์ˆ˜๋Š” $\mathbb{Z}[\sqrt{-5}]$์— ์กด์žฌํ•˜์ง€ ์•Š์Œ์„ ํ™•์ธํ–ˆ๋‹ค.

๋”ฐ๋ผ์„œ $N(\gamma)$๋Š” 1 ๋˜๋Š” 21์ธ๋ฐ, ์ด๊ฒƒ์€ โ€œeither $\gamma$ or $\delta$ is a unitโ€๋ผ๋Š” ๊ฒฐ๊ณผ๋ฅผ ์œ ๋„ํ•œ๋‹ค.

๋”ฐ๋ผ์„œ $1 + 2\sqrt{-5}$๋Š” irreducible์ด๋‹ค.

๊ฐ™์€ ๋ฐฉ๋ฒ•์œผ๋กœ $1 - 2\sqrt{-5}$๊ฐ€ irreducible์ž„๋„ ๋ณด์ผ ์ˆ˜ ์žˆ๋‹ค.


๋”ฐ๋ผ์„œ $\mathbb{Z}[\sqrt{-5}$๋Š” Integral Domain์ด์ง€๋งŒ, Unique Factorization์€ ๊ฐ€์ง€์ง€ ์•Š๋Š”๋‹ค! $\blacksquare$



์•ž์—์„œํ•œ ๋…ผ์˜๋ฅผ ํ™œ์šฉํ•˜๋ฉด, โ€œํŽ˜๋ฅด๋งˆ์˜ ๋‘ ์ œ๊ณฑ์ˆ˜ ์ •๋ฆฌโ€ (Fermatโ€™s $p = a^2 + b^2$ Theorem)์„ ์ฆ๋ช…ํ•  ์ˆ˜ ์žˆ๋‹ค!!

๋‹ค์Œ ํฌ์ŠคํŠธ: link