Proof using greatest common divisor.

AI Thread Summary
The discussion centers on proving that if c divides both a and b, then c also divides the greatest common divisor (gcd) of a and b. The proof begins by expressing a and b as multiples of c, leading to the conclusion that gcd(a, b) can be represented as a linear combination of a and b. Participants identified a redundancy in the proof regarding the substitution for d, suggesting that simplifying the argument could enhance clarity without losing validity. Overall, the proof's core idea is affirmed as correct, with suggestions for refinement to streamline the argument. The participants agree on the importance of clarity and conciseness in mathematical proofs.
bonfire09
Messages
247
Reaction score
0

Homework Statement


The problem went "If c|a and c|b then c|gcd(a,b) where cεN and a,bεZ"


Homework Equations





The Attempt at a Solution


My proof went like this " Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Lets
start with cd=ax+by where d,x,yεZ since the gcd(a,b) can be rewritten as a linear combination.
Substituting a and b we get cd=(cr)ax+(cs)ay <=> cd=c(rax+say) where rax+sayεZ. Hence
c|gcd(a,b).

is this right?
 
Physics news on Phys.org
bonfire09 said:

Homework Statement


The problem went "If c|a and c|b then c|gcd(a,b) where cεN and a,bεZ"


Homework Equations





The Attempt at a Solution


My proof went like this " Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Lets
start with cd=ax+by where d,x,yεZ since the gcd(a,b) can be rewritten as a linear combination.
Substituting a and b we get cd=(cr)ax+(cs)ay <=> cd=c(rax+say) where rax+sayεZ. Hence
c|gcd(a,b).

is this right?

Notice where I've bolded. You jumped out too far there. You're right to say that :

a=cr and b=cs

It's also good that you realize the gcd is a linear combination, but that does not automatically imply what you've stated here.

What you should say is suppose gcd(a,b) = d for some integer d. Then it follows that gcd(a,b) = d = ax + by for some integers x and y.

Now using d = ax + by the rest of your proof will follow smoothly.
 
c is a common divisor of a and b so in particular it is equal to the gcd. This part is trivial. Assume now that c is a common divisor smaller (in modulus) than gcd. You need to show that there's x integer so that gcd = x c.

Assume a = c alpha, b = c beta. What happens if (alpha,beta) =1, i.e. gcd(alpha,beta)=1 ? Who's c ?
 
I agree the first part of my proof makes a leap but not a very big leap. So my proof should read like this " Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Suppose gcd(a,b) = d for some integer d. Then it follows that gcd(a,b) = d = ax + by for some integers x and y. Thus ct=d for some tεZ which becomes ct=ax+by. Substituting a and b we get ct=(cr)ax+(cs)ay <=> ct=c(rax+say) where rax+sayεZ. Hence c|gcd(a,b).

It makes the proof look cleaner but doesn't change much. I used the fact that gcd(a,b) by definition can be written as a linear combination of some multiples of integers a and b.
 
Last edited:
bonfire09 said:
I agree the first part of my proof makes a leap but not a very big leap. So my proof should read like this " Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Suppose gcd(a,b) = d for some integer d. Then it follows that gcd(a,b) = d = ax + by for some integers x and y. Thus ct=d for some tεZ which becomes ct=ax+by. Substituting a and b we get ct=(cr)ax+(cs)ay <=> ct=c(rax+say) where rax+sayεZ. Hence c|gcd(a,b).

It makes the proof look cleaner but doesn't change much. I used the fact that gcd(a,b) by definition can be written as a linear combination of some multiples of integers a and b.

Er what I was thinking was if :
a=cr and b=cs

d = ax + by
d = crx + csy
d = c(rx + sy)

And hence c divides d. Making the substitution for d is redundant.
 
oh I realize I made an error in my original proof. "Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Suppose gcd(a,b) = d for some integer d. Then it follows that gcd(a,b) = d = ax + by for some integers x and y. Thus ct=d for some tεZ which becomes ct=ax+by. Substituting a and b we get ct=(cr)x+(cs)y <=> ct=c(rx+sy) (right here where I left a)where rx+syεZ. Hence c|gcd(a,b).

Yeah its redundant. I wouldn't say its a bad thing as it can be fixed. But I just wanted to verify that the ideas in this proof are on the right track. Ill clean this proof up.
 
Last edited:
bonfire09 said:
oh I realize I made an error in my original proof. "Let a=cr and b=cs where r,sεZ. We want to show c|gcd(a,b). Suppose gcd(a,b) = d for some integer d. Then it follows that gcd(a,b) = d = ax + by for some integers x and y. Thus ct=d for some tεZ which becomes ct=ax+by. Substituting a and b we get ct=(cr)x+(cs)y <=> ct=c(rx+sy) (right here where I left a)where rx+syεZ. Hence c|gcd(a,b).

Yeah its redundant. I wouldn't say its a bad thing as it can be fixed. But I just wanted to verify that the ideas in this proof are on the right track. Ill clean this proof up.

Yes it's the right idea for sure. Just saying why do more work than you have to right ;)?
 
Back
Top