3
$\begingroup$

I met in many places the equation

$(a^4-b^4)(c^4-d^4)=\square$

It is well known that this was investigated by Euler. But I was unable to find the general solution of this equation. Could you please clarify if there is such solution and if so could you please help to find the solution?

$\endgroup$
4
  • 1
    $\begingroup$ I may not be familiar with the field, but what does the ractangle sign mean in this context ?(for me it's a d'alembertian) $\endgroup$ Commented Apr 30, 2016 at 9:08
  • 4
    $\begingroup$ @Amir I guess it is 'a perfect square' $\endgroup$ Commented Apr 30, 2016 at 9:15
  • $\begingroup$ Yes, the square is perfect square. $\endgroup$ Commented Apr 30, 2016 at 10:52
  • $\begingroup$ There is some information about it in Tito Piezas III's A Collection of Algebraic Identities. You can google the sentence "which takes care of the first condition and modifies the second to" and a couple of sites will show up. That sentence will take you to where the equation above is mentioned. $\endgroup$ Commented Sep 30, 2025 at 19:14

2 Answers 2

1
$\begingroup$

I think general solution doesn't exist via elliptic curves and probably this can be proven rigorously.

Pick random $c,d$ where $c \ne \pm d$ and let $k=c^4-d^4$.

So we get $k(a^4-b^4)=z^2$. Dividing by $b^4$ we get $C:k(a'^4-1)=z'^2$.

This is elliptic curve and if it is of positive rank, it has infinitely many rational solutions $a'=a''/b'',z'=z''/b''$ coming from the group law. Multiplying by $b''^4$ we get the integer solutions $k(a''^4-b''^4)=z''^2 b''^2$.

On $C$ setting $a'=c/d$ gives one rational point and if it is of infinite order, it gives infinitely many solutions.

A single $k$ with positive rank gives you infinitely many integer solutions.

I suspect this happens for infinitely many $k$.


The OP asks about general polynomial solution.

Single $k$ with positive rank will give for fixed $c,d$ infinitely many solutions $a,b$ which are arbitrary large, ruling out polynomial solution.

Another approach might be to examine the K3 surface $a^4-b^4=c^4-d^4$ which gives subset of the solutions and again doesn't have complete polynomial parametrization (essentially sum of two fourth powers in two nontrivial ways, for which at least two polynomial parametrization are known).

$\endgroup$
8
  • $\begingroup$ If I understand correctly the general solution does not exist? If this can be proven does this mean that there is no polynominal solution of this equation in general? $\endgroup$ Commented Apr 30, 2016 at 10:58
  • $\begingroup$ @veg_nw Yes, I am ready to bet polynomial solution doesn't exist. $\endgroup$ Commented Apr 30, 2016 at 11:07
  • $\begingroup$ @joro You mean apart from the silly stuff like $a=kc,b=kd$? $\endgroup$ Commented Apr 30, 2016 at 11:09
  • $\begingroup$ @LevBorisov Yes. This is not complete parametrization. Though setting $a'=c/d$ gives a point on the elliptic curve and remains to show it is of infinite order. $\endgroup$ Commented Apr 30, 2016 at 11:11
  • $\begingroup$ And does this mean that there is no general "non-polynominal" solutions exist as well? $\endgroup$ Commented Apr 30, 2016 at 12:14
0
$\begingroup$

Let $n \in \mathbb{N}$ and $a_i, b_i, m \in \mathbb{Z}$ for $i=1,2,\dots,k$.
If $$ \sum_{i=1}^k a_i^n = \sum_{i=1}^k b_i^n = m$$ then $$ \left( \sum_{i=1}^k a_i^n \right) \left( \sum_{i=1}^k b_i^n \right) = m^2. $$

This Result can be further Generalized to cubes and higher powers(product of $x$ number of Sums is equal to $m^x$) .

If $$m=a^n-b^n=c^n-d^n$$ for $n \in \{1,2,3,4\}$ and $a,b,c,d \in \mathbb{N}$ then $$(a^n-b^n)(c^n-d^n)=m^2$$

Parametrizations for $n=2$ :

$1$st parametrization :

Take $p=ac+bd$, $q=ac-bd$, $r=ad+bc$ and $s=ad-bc$

Wkt $p^2-q^2=r^2-s^2=4abcd$

Substitute it for above result then we get

$(p^2-q^2)(r^2-s^2)=(4abcd)^2$

$2$nd parametrization :

Take $p=ac+bd$ , $q=ad+bc$ , $r=ad-bc$ and $s=ac-bd$

Wkt $p^2-q^2=r^2-s^2=(a^2-b^2)(c^2-d^2)$

Substitute it for above result then we get

$(p^2-q^2)(r^2-s^2)=(a^2-b^2)^2(c^2-d^2)^2$

Similarly for $n=3$ and $n=4$ we can obtain parametrizations.

Some examples :

For $n=2$ :

Wkt $4^2+7^2=1^2+8^2$

$(4^2-1^2)(8^2-7^2)=15^2$

$(7^2-1^2)(8^2-1^2)=48^2$

For $n=3$ :

Wkt $10^3+9^3=1^3+12^3$

$(9^3-1^3)(12^3-10^3)=(6^3+8^3)^2=728^2$

$(10^3-1^3)(12^3-9^3)=999^2$

For $n=4$ :

Wkt $133^4+134^4=59^4+158^4$

$(133^4-59^4)(158^4-134^4)=300783360^2$

$(134^4-59^4)(158^4-133^4)=310300575^2$

$\endgroup$
2
  • $\begingroup$ What is "Wkt"?? $\endgroup$ Commented Sep 30, 2025 at 22:34
  • 1
    $\begingroup$ @GerryMyerson , "Wkt" means We know that. $\endgroup$ Commented Sep 30, 2025 at 23:07

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.