1

Consider the following Diophantine equation. $$5(p^2+q^2+r^2+s^2+t^2)^2-7(p^4+q^4+r^4+s^4+t^4)=90pqrst$$ This equation was discussed at The Diophantine equation $5(p^2+q^2+r^2+s^2+t^2)^2 - 7(p^4+q^4+r^4+s^4+t^4) = 90pqrst$

The author was interested about the source, not the solution. I found the problem quite interesting. I wrote a code which generated all the solutions $1 \leq p \leq q \leq r \leq s \leq t \leq 500$.

p q r s t
1 1 1 1 1
1 1 1 1 2
1 1 1 2 2
1 1 1 2 4
1 1 2 3 3
1 1 2 3 6
1 2 4 7 7
1 2 7 7 17
1 2 34 47 127
1 17 42 63 124
2 7 19 47 59
2 13 39 97 99
3 3 6 42 87
3 3 27 87 189
6 21 138 417 483
11 11 11 154 286

According to the magazine mentioned in the link, this equation has a relation to the Markov numbers (https://en.wikipedia.org/wiki/Markov_number).

I wonder if there are similar tree like structure for the solutions of this problem.

piepie
  • 547
  • The "tree structure" of the Markoff equation comes from the fact that it is quadratic in each variable. – yoyo Dec 18 '20 at 01:45
  • 1
    Recommend trying $p^2 + q^2 + r^2 + s^2 + t^2 = 5pqrst$ which does have jumping behavior. You may also enjoy $p^2 + q^2 + r^2 + s^2 + t^2 = 4pqrst$ and $p^2 + q^2 + r^2 + s^2 + t^2 = pqrst$ which have jumping behavior but do not include $(1,1,1,1,1),$ rather $(2,1,1,1,1),$ for the first and $(4,3,3,1,1),$ for the second. – Will Jagy Dec 18 '20 at 21:38

1 Answers1

1

The answer depends on the precise meaning of your "similar tree like structure". Consider a very general situation. Fix a symmetric polynomial $\,P(x_1,\dots,x_n).\,$ Define $$ T:=\{(x_1,\dots,x_n)\in\mathbb{Z}^n:P(x_1,\dots,x_n)=0\}.\tag{1}$$ Define a graph $\,G\,$ where the vertices $\,V\,$ are equivalence classes of $\,T\,$ where two elements of $\,T\,$ are regarded as equivalent if they are permutations of each other. Define two vertices as adjacent if they are equal up to permutation except for at most one component. Because $\,P\,$ is a polynomial, there can be at most $\,n-1\,$ vertices adjacent to a given vertex. If there are no cycles, then the graph is a tree. In the case of the Markov numbers, $\,P(x,y,z) = x^2+y^2+z^2-3xyz.\,$ In the case in your question, $$ P(p,q,r,s,t)=5(p^2+q^2+r^2+s^2+t^2)^2-7(p^4+q^4+r^4+s^4+t^4)-90pqrst. \tag{2}$$ It is probably true that there are no cycles but I have no proof for that, even if the numbers are restricted to positive integers. Thus, the solutions form a tree like structure.

Somos
  • 35,251
  • 3
  • 30
  • 76
  • Thanks for the general idea. Markoff-like equations for more than three variables or higher order is not well understood. you can find this equation in Hirzebruch and Zagier, "The Atiyah-Singer theorem and elementary number theory" (pg. 159, eqn. (7)), as mentioned by @yoyo in the original link. – piepie Dec 18 '20 at 03:59