Skip to content

Latest commit

 

History

History
188 lines (148 loc) · 7.3 KB

File metadata and controls

188 lines (148 loc) · 7.3 KB
title author header-includes abstract
Abstract Algebra by Pinter, Chapter 23, question B3
Amir Taaki
- \usepackage{mathrsfs} - \usepackage{mathtools} - \usepackage{extpfeil} - \DeclareMathOperator\ker{ker} - \DeclareMathOperator\ord{ord} - \DeclareMathOperator\gcd{gcd} - \DeclareMathOperator\lcm{lcm} - \DeclareMathOperator\mod{mod} - \DeclareMathOperator\char{char}
Chapter 23 on Number Theory

Proof

Initial Question

We are given $k$ congruences $$x \equiv c_1 (\mod m_1) \qquad x \equiv c_2 (\mod m_2) \qquad \cdots \qquad x \equiv (\mod m_k)$$ for all $i, j \in {1, \dots, k}$ $$c_i \equiv c_j (\mod d_{ij})$$ where $d_{ij} = \gcd(m_i, m_j)$.

Prove there is an $x$ satisfying all $k$ congruences simultaneously, and the solution is of the form $$x \equiv c (\mod t)$$ where $t = \lcm(m_1, m_2, \dots, m_k)$.

Simultaneous Solution for Three Elements

We will proceed to prove these statements through induction, first starting with the case of proving there is a simultaneous solution for $c_1, c_2$ and $c_3$.

It has been shown earlier in theorem 3 that there is a solution for two equations $x \equiv a (\mod n)$ and $x \equiv b (\mod m)$, only exists if $$a \equiv b (\mod d)$$ $$d = \gcd(m, n)$$

For the first two equations, there is therefore a simultaneous solution because $$c_1 \equiv c_2 (\mod d_{12})$$

Earlier in theorem 4, it was shown that if $x \equiv a (\mod n)$ and $x \equiv b (\mod m)$ have a simultaneous solution, it is of the form $$x \equiv c (\mod t)$$ $$t = \lcm(m, n)$$

So therefore the solution of $x \equiv c_1 (\mod m_1)$ and $x \equiv c_2 (\mod m_2)$ is $$ x \equiv c (\mod t)$$ $$t = \lcm(m_1, m_2)$$

We want to know if there is a solution $x$ for $x = c (\mod t)$ and $x = c_3 (\mod m_3)$. That is whether the statement $$c_3 \equiv c [\mod \gcd(t, m_3)]$$ is true.

But we know that $\gcd(a, \lcm(b, c)) = \lcm(\gcd(a, b), \gcd(a, c))$ so \begin{align*} \gcd(t = \lcm(m_1, m_2), m_3) &= \lcm(\gcd(m_1, m_3), \gcd(m_2, m_3)) \ &= \lcm(d_{13}, d_{23}) \end{align*}

So we want to know whether this statement is true \begin{align*} c_3 &\equiv c [\mod \gcd(t, m_3)] \ &\equiv c [\lcm(d_{13}, d_{23})] \end{align*}

At the start it was stated that $c_3 \equiv c_1 (\mod d_{13})$, and we also we know that $$c \equiv c_1 (\mod m_1) \implies c \equiv c_1 (\mod d_{13})$$ $$\therefore c \equiv c_3 (\mod d_{13})$$ Likewise $c_3 \equiv c_2 (\mod d_{23}) \implies c \equiv c_3 (\mod d_{23})$

Now from the last part of theorem 4, we note that $$m \mid (x - c) \text{ and } n \mid (x - c) \iff t \mid (x - c)$$ or $$x \equiv c (\mod m) \text { and } x \equiv c (\mod n) \iff x \equiv c (\mod t)$$

Note that $$d_{13} \mid (c - c_3) \text { and } d_{23} \mid (c - c_3) \iff lcm(d_{13}, d_{23}) \mid (c - c_3)$$ or $$c \equiv c_3 (\mod d_{13}) \text{ and } c \equiv c_3 (\mod d_{23}) \iff c \equiv c_3 [\mod \lcm(d_{13}, d_{23})]$$

That is we can state that $$c_3 \equiv c [\mod \lcm(d_{13}, d_{23})]$$ But $\lcm(d_{13}, d_{23}) = \gcd(t, m_3)$. So by theorem 3 because $$c_3 \equiv c [\mod \gcd(t, m_3)]$$ there is a simultaneous solution of \begin{align*} x &\equiv c (\mod t) \ x &\equiv c_3 (\mod m_3) \end{align*} And this is also the solution for \begin{align*} x &\equiv c_1 (\mod m_1) \ x &\equiv c_2 (\mod m_2) \end{align*}

Generalizing to $k + 1$ through induction

Now we will generalize this using induction on $k + 1 \in \mathbb{Z}$ terms where we assume $S_k$ is true, proving the statement $S_{k + 1}$ is true, and therefore it is true for all integers.

Assume there is a solution of $k$ congruences $$x \equiv c_1 (\mod m_1) \qquad \cdots \qquad x \equiv c_k (\mod m_k)$$ of the form $$x \equiv c (\mod t)$$ $$t = \lcm(m_1, \dots, m_k)$$

Note that $\forall i, j \in {1, \dots, k}$ $$c_i \equiv c_j (\mod d_{ij})$$ $$d_{ij} = \gcd(m_i, m_j)$$ that is $$c_{k + 1} = c_i (\mod d_{k + 1, i})$$

We want to know if there an $x (\mod t')$ which is the solution for $x \equiv c (\mod t)$ and $x \equiv c_{k + 1} (mod m_{k + 1})$. That is whether the statement $$c_{k + 1} \equiv c [\mod \gcd(t, m_{k + 1})]$$ is true or not.

Relation between gcd and lcm operators

From chapter 22, exercise H4, let $a \star b = \gcd(a, b)$ and $a \circ b = \lcm(a, b)$ then it is trivial to show that $$a \star (b \circ c) = (a \star b) \circ (a \star c)$$ and we know that the $\lcm$ operation is associative. $$m_1 \circ m_2 \circ \cdots \circ m_k = m_1 \circ (m_2 \circ (\cdots \circ m_k))$$ so \begin{align*} m_{k + 1} \star (m_1 \circ m_2 \circ \cdots \circ m_k) &= (m_{k + 1} \star m_1) \circ (m_{k + 1} \star (m_2 \circ \cdots \circ m_k)) \ &= (m_{k + 1} \star m_1) \circ (m_{k + 1} \star m_2) \circ (m_{k + 1} \star (m_3 \circ \cdots \circ m_k)) \ &= (m_{k + 1} \star m_1) \circ \cdots \circ (m_{k + 1} \star m_k) \end{align*} That is $$\gcd(\lcm(m_1, \dots, m_k), m_{k + 1}) = \lcm(\gcd(m_1, m_{k + 1}), \dots, \gcd(m_k, m_{k + 1}))$$

Proving equivalency holds under gcd for $k + 1$

At the beginning it was stated that $\forall i \in {1, \dots, k}$ $$c_{k + 1} \equiv c_i (\mod d_{k + 1, i})$$ and we also know that $$c \equiv c_i (\mod m_i)$$ $$c - c_i = q m_i = q(sd_{k + 1, i})$$ $$\implies c \equiv c_i (\mod d_{k + 1, i})$$

Generalizing lcm to Multiple Arguments

The $\lcm$ is defined as if $c = \lcm(a, b)$ then

  1. $a \mid c$ and $b \mid c$
  2. For any $x$ if $a \mid x$ and $b \mid x \implies c \mid x$

This can be generalized for any number of arguments in the $\lcm$ by noting that since $c = \lcm(x_1, x_2, \dots, x_n)$ then $\forall i \in {1, \dots, n}$ then 1. $x_i \mid c$ for 2., note that the common multiples of ${ x_1, \dots, x_n}$ form an ideal of $\mathbb{Z}$ by $\langle c \rangle = \langle x \rangle \cap \cdots \cap \langle x_n \rangle$, and so every common multiple is a multiple of $c$.

$\therefore$ any $v$ such that $\forall x_i \in X: x_i \mid v \implies c \mid v$.

Solution $c \equiv c_i$ is also a Solution in the lcm of the gcds

From theorem 4, we generalize that $$m_1 \mid x, \cdots, m_n \mid x \implies t \mid x$$ $$m_1 \mid (x - c), \cdots, m_n \mid (x - c) \implies t \mid (x - c)$$ $$x \equiv c (\mod m_1) \qquad \cdots \qquad x \equiv c (\mod m_n) \implies x \equiv c (\mod t)$$ where $t = \lcm(m_1, \dots, m_n)$

Now note that $$d_{k + 1, 1} \mid (c - c_i) \qquad \cdots \qquad d_{k + 1, k} \mid (c - c_i) \implies \lcm(d_{k + 1, 1}, \dots, d_{k + 1, k}) \mid (c - c_i)$$ or $$c \equiv c_i (\mod d_{k + 1, 1}) \qquad \cdots \qquad c \equiv c_i (\mod d_{k + 1, k}) \implies c \equiv c_i [\mod \lcm(d_{k + 1, 1}, \dots, d_{k + 1, k})]$$

There is a Common Solution for $c$ and $c_{k + 1}$

So, $$c \equiv c_i [\mod \lcm(\gcd(m_{k + 1}, m_1), \dots, \gcd(m_{k + 1}, m_k))]$$

But we know that $$\lcm(\gcd(m_{k + 1}, m_1), \dots, \gcd(m_{k + 1}, m_k)) = \gcd(\lcm(m_1, \dots, m_k), m_{k + 1})$$ $$\implies c \equiv c_i [\mod \gcd(t, m_{k + 1})]$$ where $t = \lcm(m_1, \dots, m_k)$

And because of this, by theorem 3, because $\forall i \in {1, \dots, k}, c \equiv c_i [\mod \gcd(t, m_{k + 1})]$, there is an $x$ such that $$x \equiv c (\mod t)$$ $$x \equiv c_{k + 1} (\mod m_{k + 1})$$ which because $x \equiv c (\mod t)$, this is also the solution for $$x \equiv c_1 (\mod m_1)$$ $$\cdots$$ $$x \equiv c_k (\mod m_k)$$

Furthermore this solution takes the form \begin{align*} x &\equiv c' [\mod \lcm(t, m_{k + 1})] \ &\equiv c' (\mod t') \end{align*} where $t' = \lcm(m_1, m_2, \dots, m_k)$