1

In my algebra class I was given as homework assignment to prove that:

Given an integral domain $A$ and $a,b,c,d,e \in A$. Then if $d = \gcd(b,c)$ and $e = \gcd(ac, ab)$ then $e = ad$.

It is easy to see that $ad \mid ab$ and $ad \mid ac$, so this implies that $\exists q \in A; e = q(ad)$. Now I'm having problems showing that $q = 1_A$. Working through some equalities, I proved that $\forall n\in \mathbb{N}, q^n \mid d$ so this gives me the light suspicion that I might be going the right way, but I'm pretty positive that this does not imply $q = 1_A$.

Does anyone want to give me a hint?

Thanks in advance :)

Bernard
  • 175,478
rie
  • 833
  • 6
  • 13
  • 2
    There are integral domains in which there are no g.c.d. s. Yoy need a gcd domain( (for instance a UFD or a PID). – Bernard Apr 02 '16 at 21:13
  • 3
    @Bernard I am aware of that, but in this case the problem states that the element $e$ is indeed a gcd of $ac$ and $ab$. There's no problem with the extistence there – rie Apr 02 '16 at 22:54
  • See the 2nd proof here in the linked dupe. – Bill Dubuque Jan 05 '23 at 10:41

1 Answers1

2

Here's a solution:

Set $\;b=b'd$, $\;c=c'd\;\;$ ($b'\wedge c'=1$), $\;ab=\beta e$, $\;ac=\gamma e\;\;$ ($\beta\wedge\gamma=1$). We have: \begin{align*} q\beta e&=qab=qadb'=eb',\\ q\gamma e& =qac=qadc'=ec', \end{align*} whence, since $A$ is a domain, $\;b'=q\beta,\enspace c'=q\gamma$. As $b'\wedge c'=1$, these relations imply $q$ is a unit.

Bernard
  • 175,478