Arithmetic Homework: Showing Bezout Theorem

geoffrey159
Messages
535
Reaction score
72

Homework Statement


We are given ##a_1,...,a_n ## in ## \mathbb{N}^\star ##, all mutually prime, and ## a = a_1 \times ... \times a_n ##.

Show that for all ##(b_1,...,b_n)\in\mathbb{Z}^n##, there is ##\beta \in \mathbb{Z} ## such that for all ##x \in \mathbb{Z} ## :

## (\forall i = 1 ... n, \ x \equiv b_i (\text{mod } a_i) ) \iff (x\equiv \beta \ (\text{mod } a))##

Homework Equations


Bezout theorem

The Attempt at a Solution



Please could you tell me if this is correct ? Thanks !

The idea is that for all ##i = 1... n##, ##a_i## and ## a/a_i ## are mutually prime, so that there are ##(u_i,v_i)\in \mathbb{Z}^2## such that ## 1 = a_i u_i + \frac{a}{a_i} v_i ##. I can now write ## b_i = a_i b_i u_i + \frac{a}{a_i} b_i v_i ##

(##\Rightarrow##) :
We have : ## a_i | ( x - b_i) \Rightarrow a_i | (x - a_i b_i u_i - \frac{a}{a_i} b_i v_i ) \Rightarrow a_i | (x - \frac{a}{a_i} b_i v_i) ##.
Also, for ## i\neq j##, we have that ## a_i | \frac{a}{a_j} b_j v_j \Rightarrow a_i | \sum_{j\neq i} \frac{a}{a_j} b_j v_j ##

Setting ##\beta = \sum_{ j} \frac{a}{a_j} b_j v_j ##, we have that ##a_i | (x - \beta) ##. Since the ##a_i## are mutually prime, ## a| (x-\beta) ##

(##\Leftarrow##) :
##a## is the least common multiple of ##(a_1,...,a_n)## (because they are mutually prime).
So for all ## i = 1...n##, ## a_i | (x-\beta) ##.
Also, reusing the fact that ## a_i | \sum_{j\neq i} \frac{a}{a_j} b_j v_j ##, we have that ##a_i | (x - \frac{a}{a_i} b_i v_i) = x - b_i + a_i b_i u_i##.
Therefore ## a_i | ( x - b_i) ##
 
Physics news on Phys.org
The second part of the question is :
' Show that ## \mathbb{Z}/a\mathbb{Z} ## and ## \mathbb{Z}/a_1\mathbb{Z} \times ... \times \mathbb{Z}/a_n\mathbb{Z} ## are isomorphic. Use ##f(\text{cl}_a(x)) = (\text{cl}_{a_1}(x),...\text{cl}_{a_n}(x))## '

where ##\text{cl}_u(x) = \{ y\in \mathbb{Z} , u | y-x \} ####\bullet## Injectivity of ##f##:
## f(\text{cl}_a(x)) = f(\text{cl}_a(x')) \Rightarrow \forall i = 1...n \ \text{cl}_{a_i}(x) = \text{cl}_{a_i}(x') \Rightarrow \forall i = 1...n \ a_i | x-x' ##.
Since the ##a_i## are mutually prime, ## a | x - x' ## and therefore ##\text{cl}_a(x) = \text{cl}_a(x')##

##\bullet## Surjectivity of ##f##:
Let ##y_i \in \mathbb{Z}/a_i\mathbb{Z}##, for all ## i = 1...n##. There are ##x_i,...,x_n \in \mathbb{Z} ## such that ## y_i = \text{cl}_{a_i}(x_i)##.
We are looking for ##x## such that for all ##i = 1...n##, ## \text{cl}_{a_i}(x_i) = \text{cl}_{a_i}(x) ##. So we must find ##x## that solves ## a_i | x - x_i ## for all ##i = 1...n##. According to question 1, it happens if and only if ## a | x - \beta ##. So ##x = \beta ## works and ##f(\text{cl}_a(\beta)) = (\text{cl}_{a_1}(x_1),...\text{cl}_{a_n}(x_n)) = (y_1,...,y_n) ##

##\bullet## ##f## is a morphism from the group ##(\mathbb{Z}/a\mathbb{Z},+)## to ## (\mathbb{Z}/a_1\mathbb{Z} \times ... \times \mathbb{Z}/a_n\mathbb{Z},+)##, we just need to use the fact that ## \text{cl}_u(x) + \text{cl}_u(x') = \text{cl}_u(x+x') ##

Are you ok with that ?
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...

Similar threads

Replies
13
Views
1K
Replies
1
Views
1K
Replies
14
Views
2K
Replies
3
Views
3K
Replies
1
Views
2K
Replies
5
Views
2K
Replies
2
Views
1K
Back
Top