2
$\begingroup$

Show that if $x,y,z$ are positive integers, then $(xy + 1)(yz + 1)(zx + 1)$ is a perfect square if and only if $xy +1, yz +1, zx+1$ are all perfect squares. Find infinitely many triples (a,b,c) of positive integers such that a, b, c are in arithmetic progression and such that $ab + 1, bc +1, ca+1$ are perfect squares.

First of all I tried to prove that $xy +1, yz +1, zx+1$ were relatively prime. I didn't succeed.

$$\gcd(xy+1, yz+1) = \gcd(y(x-z)+1,yz+1)$$

For the second problem, I have discovered the solution $(2,4,12)$ which isn't in arithmetic progression.

I have tried googling it but didn't find an easy and satisfactory solution.

Any help would be appreciated.

$\endgroup$
  • $\begingroup$ @labbhattacharjee My question also includes another question to find them. It's not a duplicate. The other question doesn't include this in case you haven't noticed. $\endgroup$ – TheRandomGuy Mar 19 '16 at 10:15
  • $\begingroup$ @Macavity The first question has not been properly answered and what are Pell equations? $\endgroup$ – TheRandomGuy Mar 19 '16 at 11:52
4
$\begingroup$

First question is a well known problem studied by Euler and some variants of it still out of our reach . For every positive integers $(x,y,z)$ let $P(x,y,z)$ denote the predicate: $(xy+1)(yz+1)(zx+1)$ is a square and not all of $(xy+1)$, $(yz+1)$,$(zx+1)$ are squares

Proof by infinite descent:

we well prove by the infinite descent that if there exists a solution $(x,y,z)$ such that $P(x,y,z)$ is true than there exists another solution $(p,q,r)$ such that $P(p,q,r)$ is true and $p+q+r$ less than $x+y+z$.

Let $(x,y,z)$ be tuple of positive integers such that $P(x,y,z)$ is true and $x\leq y \leq z$, consider the two positive integers $s_{+}$ and $s_{-}$ defined by : $$s_{\mp}=x+y+z+2xyz\mp\sqrt{(xy+1)(yz+1)(zx+1)} \tag 1$$ This integers verify : $$x^2+y^2+z^2+s_{\mp}^2-2(xy+yz+zx+xs_{\mp}+ys_{\mp}+zs_{\mp})-4xyzs_{\mp}-4=0 \tag2$$ And we can also prove the following important identities (they are basically the same): $$ \begin{align} (x+y-z-s_{\mp})^2&&=&& 4(xy+1)(zs_{\mp}+1) \tag 3 \\ (x+z-y-s_{\mp})^2&&=&& 4(xz+1)(ys_{\mp}+1) \tag 4 \\ (x+s_{\mp}-z-y)^2&&=&& 4(yz+1)(xs_{\mp}+1) \tag 5 \end{align}$$ we can use this identities to proof that $P(x,y,s_{\mp})$ holds, by multiplying $(4)$ and $(5)$ you will get that $(xy+1)(ys_{\mp}+1)(xs_{\mp}+1) $ is a square and not all of $(xy+1)$,$(xs_{\mp}+1)$ , $(ys_{\mp}+1)$ are square (as we have $(xz+1)$ is a square iff $xs_{\mp}+1)$ is a square and $(yz+1)$ is a square iff $(ys_{\mp}+1)$ is a square using $(4)$ and $(5)$).

Now the important par is to prove that either $x+y+s_{+}<x+y+z$ or $x+y+s_{-}<x+y+z $ this is equivalent tp proving that $s_{-}s_{+}<z^2$ which is true because : $$s_{-}s_{+}=x^2+y^2+z^2-2(xy+yz)-4=z^2-x(2z-x)-y(2z-y)-4<z^2 $$ (remember that $x\leq y \leq z$)

Reference: When Is (xy + 1)(yz + 1)(zx + 1) a Square? Kiran S. Kedlaya Mathematics Magazine Vol. 71, No. 1 (Feb., 1998), pp. 61-63

Second Question: The following Pell equation have infinitely many solutions : $$x^2-3y^2=1 $$

For any solution $(n,m)$ of this Pell equation, let $(a,b,c)=(2n-m,2n,2n+m)$ then $ab+1,bc+1,ca+1$ are all squares.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.