3

Is there any positive integer solutions of $x$, $y$, $z$ such that both $$\frac{x^2+y^2+z^2}{x+y+z}$$ and $$\frac{x^2+y^2+z^2}{xyz}$$ are integers?

Edit: Forgot to mention that $x$, $y$, $z$ are distinct positive integers.

Ng Chung Tak
  • 18,990
user3137471
  • 169
  • 8

2 Answers2

1

A remark, how to start: The second condition can be written as Diophantine equation $$ x^2+y^2+z^2=nxyz, $$ and it is easy to see that we must have $n=1$ or $n=3$. This is already very helpful. Compare also with this question.

For $n=3$ the triples $(x,y,z)$ with $\frac{x^2+y^2+z^2}{xyz}=3$ are known as Markov numbers, and we can produce a Markov tree. Now it gets easier if we also require that $(x,y,z)$ is a solution of the Diophantine equation $x^2+y^2+z^2=m(x+y+z)$.

Dietrich Burde
  • 130,978
  • reference I like http://zakuski.utsa.edu/~jagy/Hurwitz_A_1907.pdf – Will Jagy Aug 29 '16 at 17:59
  • Thanks Dietrich, it is really helpful and pushed me to the right direction. I will try to see if along the Markov Tree is there any solution possible for Diophantine equation $$x^2+y^2+z^2=m(x+y+z)$$ Also thanks for the Reference Will Jagy. – user3137471 Aug 29 '16 at 18:04
  • 1
    @WillJagy Yes, definitely! Also, I like it to read in German.... – Dietrich Burde Aug 29 '16 at 18:10
  • Thanks, I think this is the closest we can get. Tried first million members of markov chain without any success. – user3137471 Aug 31 '16 at 11:23
-1

Special case

Well, if $x = y = z$ then your equations become

$$\frac{3x^2}{3x} = x$$

$$\frac{3x^2}{x^3} = \frac{3}{x}$$

If they are coupled, then you notice that $x = 1$ is the simplest result. There is also $x = 3$: $3/x$ is an integer if and only if $x = 1$ or $x = 3$.

Unless you count negative integers too, in that case you have also $x = y = z = -1$ and $x = y = z = -3$

Enrico M.
  • 26,114