Prove that the congruences admit a simultaneous solution

  • Thread starter Thread starter Math100
  • Start date Start date
Math100
Messages
823
Reaction score
234
Homework Statement
Prove that the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ## admit a simultaneous solution if and only if ## gcd(n, m)\mid (a-b); ## if a solution exists, confirm that it is unique modulo ## lcm(n, m) ##.
Relevant Equations
None.
Proof:

Consider the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ##.
Suppose ## \exists ## a solution for ## x ##.
Let ## d=gcd(n, m) ##.
This means ## \exists r, s\in\mathbb{Z} ## such that ## n=dr ## and ## m=ds ##.
Then ## x\equiv a\pmod {n}\implies x=a+nt, t\in\mathbb{Z}, x\equiv b\pmod {m}\implies x=b+mk, k\in\mathbb{Z} ##.
Now we have ## a+nt=b+mk\implies nt-mk=b-a ##.
Observe that substituting for ## m, n ## produces:
## d(sk-rt)=a-b ##.
Thus, ## d=gcd(n, m)\mid (a-b) ##.
Conversely, suppose ## gcd(n, m)\mid (a-b) ##.
Then ## d=gcd(m, n) ## and ## d\mid (a-b) ##.
This means ## dt=a-b ## for some ## t\in\mathbb{Z} ## such that ## d=nx_{0}+my_{0} ##.
Observe that ## dt=nx_{0}t+my_{0}t=a-b\implies my_{0}t+b=a-nx_{0}t ##.
Let ## x\equiv a\pmod {n}, x\equiv b\pmod {m} ## and ## y ## be any other solution.
Since ## y\equiv a\pmod {n} ## and ## y\equiv b\pmod {m} ##,
it follows that ## x\equiv y\pmod {n} ## and ## x\equiv y\pmod {m} ##.
Thus, ## \exists ## a simultaneous solution.
Therefore, the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ## admit a simultaneous solution
if and only if ## gcd(n, m)\mid (a-b); ## and a solution exists, hence, ## x\equiv y\pmod {lcm(m, n)} ##.
 
Physics news on Phys.org
Math100 said:
Homework Statement:: Prove that the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ## admit a simultaneous solution if and only if ## gcd(n, m)\mid (a-b); ## if a solution exists, confirm that it is unique modulo ## lcm(n, m) ##.
Relevant Equations:: None.

Proof:

Consider the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ##.
Suppose ## \exists ## a solution for ## x ##.
Let ## d=gcd(n, m) ##.
This means ## \exists r, s\in\mathbb{Z} ## such that ## n=dr ## and ## m=ds ##.
Then ## x\equiv a\pmod {n}\implies x=a+nt, t\in\mathbb{Z}, x\equiv b\pmod {m}\implies x=b+mk, k\in\mathbb{Z} ##.
Now we have ## a+nt=b+mk\implies nt-mk=b-a ##.
Observe that substituting for ## m, n ## produces:
## d(sk-rt)=a-b ##.
Thus, ## d=gcd(n, m)\mid (a-b) ##.
Conversely, suppose ## gcd(n, m)\mid (a-b) ##.
Then ## d=gcd(m, n) ## and ## d\mid (a-b) ##.
This means ## dt=a-b ## for some ## t\in\mathbb{Z} ## such that ## d=nx_{0}+my_{0} ##.
Observe that ## dt=nx_{0}t+my_{0}t=a-b\implies my_{0}t+b=a-nx_{0}t ##.
Let ## x\equiv a\pmod {n}, x\equiv b\pmod {m} ## and ## y ## be any other solution.
Since ## y\equiv a\pmod {n} ## and ## y\equiv b\pmod {m} ##,
it follows that ## x\equiv y\pmod {n} ## and ## x\equiv y\pmod {m} ##.
Thus, ## \exists ## a simultaneous solution.
Therefore, the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ## admit a simultaneous solution
if and only if ## gcd(n, m)\mid (a-b); ## and a solution exists, hence, ## x\equiv y\pmod {lcm(m, n)} ##.
The equivalence proof is correct, a bit confusing, but correct. I am a little bit lost in the uniqueness part.

Let me suggest a hopefully slimmer version:

Consider the congruences ## x\equiv a\pmod {n} ## and ## x\equiv b\pmod {m} ##.

(a) Let ##x## be such a solution.
Then ##n\,|\,(x-a)## and ##m\,|\,(x-b).##
Moreover, ##\operatorname{gcd}(n,m)\,|\,n \wedge \operatorname{gcd}(n,m\,|\,m.##
Thus ##\operatorname{gcd}(n,m)\,|\,(x-a)\wedge \operatorname{gcd}(n,m)\,|\,(x-b).##
Finally, ##\operatorname{gcd}(n,m)\,|\,[(x-b)-(x-a)]=a-b.##

It follows your proof, only without the many parameters that are necessary to establish equations. The divisibility properties are sufficient.

(b) Conversely, suppose ## \operatorname{gcd}(n, m)\,|\, (a-b), ## i.e. ##\operatorname{gcd}(n, m)\cdot t=a-b## for some ##t\in \mathbb{Z}.##
Bézout's identity guarantees the existence of ##x_0,y_0\in \mathbb{Z}## such that ## \operatorname{gcd}(n, m)=x_0n+y_0m.## Hence
\begin{align*}
x&:=a-nx_0t=\operatorname{gcd}(n,m)\cdot t+b - nx_0t\\&\phantom{:}
=\operatorname{gcd}(n,m)\cdot t +b - \operatorname{gcd}(n,m)\cdot t +y_0m\cdot t=b +my_0t
\end{align*}
is a solution to ##x\equiv a \pmod n## and ##x\equiv b \pmod m## what had to be shown.

This uses again your ideas, only that I name Bézout's identity and save a few parameters. It is easier to read with fewer variables.

(c) Let ##x,y## be two solutions.
Then ##x-y \equiv 0=a-a \pmod n## and ##x-y \equiv 0=b-b \pmod m.## Thus ##n\,|\,(x-y)## and ##m\,|\,(x-y)##. Therefore ##\operatorname{lcm}(n,m)\,|\,(x-y)## or ##x=y+ s\cdot \operatorname{lcm}(n,m).##

Maybe this was what you meant. I just wasn't sure.
 
  • Like
Likes   Reactions: berkeman and Math100
Isn't this given by the Chinese Remainder Theorem? You only need to prove the remainder is Chinese ;).
 
WWGD said:
Isn't this given by the Chinese Remainder Theorem? You only need to prove the remainder is Chinese ;).
Yes, it is. But the question asked for a proof.
 
fresh_42 said:
Yes, it is. But the question asked for a proof.
Doesn't CRT provide conditions on the uniqueness of solutions?
 
WWGD said:
Doesn't CRT provide conditions on the uniqueness of solutions?
According to Wikipedia, the above exercise is basically the CRT, not the algorithmic version, but the theoretical.
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
Replies
2
Views
2K
Replies
1
Views
1K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 6 ·
Replies
6
Views
1K