1

If $gcd(a,b) = 1$ then is it true that $gcd(2a,2b) = 2$? More generally, if $gcd(a,b) = k$ then is it true that $gcd(an,bn) = n*gcd(a,b) = kn$?

Looking at Bézout's identity this seems to be the case since we could simply multiply both sides of the equation $ax + by = d$ by the constant $n$.

terminex9
  • 455

2 Answers2

2

It is special case of the gcd distributive law $\ \gcd(na,nb)\, =\, n\gcd(a,b).\ $

See here for a few proofs of this law (by Bezout, universal property, and prime factorizations).

Community
  • 1
Bill Dubuque
  • 272,048
1

Let $n\ne 0$, and let $k=\gcd(a,b)$. It is not hard to see that $nk$ divides each of $na$ and $nb$.

Now we need to show that every common divisor of $na$ and $nb$ divides $nk$. Here we use the Bezout "Identity." There exist integers $s$ and $t$ such that $sa+tb=k$. Multiply through by $n$. We get $s(na)+t(nb)=nk$. From this we can see that any common divisor of $na$ and $nb$ divides $nk$.

Remark: The harder part of the argument was already in the OP.

André Nicolas
  • 507,029
  • This is the Bezout method I mention. In the link I highlight how it arises from the preservation under scalings of objects defined by linear equations and divisibility relations, i.e. by innate scaling symmetry. – Bill Dubuque Feb 12 '15 at 04:35
  • The link, and your earlier posts along similar lines, are very useful. – André Nicolas Feb 12 '15 at 04:37