3

I. Near-perfect Euler bricks

Given System 1,

\begin{align} a^2 + b^2 &= \square_1\\[4pt] a^2 + c^2 &= \square_2\\[4pt] b^2 + c^2 &= \square_3 + \color{red}{h^2}\\[4pt] a^2 + b^2 + c^2 &= \square_4 + \color{red}{h^2}\end{align}

If $h=0$, then this is the perfect Euler brick, though none have been found. System 1 is equivalent to System 2,

\begin{align} a^2 + b^2 &= \square_a\\[4pt] a^2 + c^2 &= \square_b\\[4pt] a^2 + d^2 &= \square_c\\[4pt] b^2 + c^2 - d^2 &= h^2\end{align}

In this post, Tomita (after some modifications I made) found a solution $(a,b,c,d)$ to the first three equations as,

\begin{align} a &= 2(n^2-1)\\[3pt] b &= 3(n^2-1)/2\\[3pt] c &= \frac{(n^2-1)(z^2-1)}z\\[3pt] d &= \frac{(n^2-1)^2-z^2}z\end{align}

where the fourth becomes a quartic in $z$ to be made a square,

$$\frac{4 n^2 (n^2 - 2) z^4 + 9 (n^2 - 1)^2 z^2 - 4 n^2 (n^2 - 2) (n^2 - 1)^2}{4z^2} = h^2$$

I found this has the simple initial polynomial solution,

$$z_1 = \frac{n^2-1}n\qquad$$


II. Question

Q: Since the quartic in $z$ is birationally equivalent to an elliptic curve, then there are infinitely many rational $z_n$. What is the simplest polynomial $z_2, z_3,$ and so on?

1 Answers1

4

Hence, we consider the quartic equation, $$ v^2 = 4n^2(n^2-2)z^4+9(n^2-1)^2z^2-4n^2(n^2-2)(n^2-1)^2$$ We know the initial solution $P(z,v)=( \dfrac{n^2-1}{n}, \dfrac{(n-1)(n+1)(n^2+1)}{n}).$

$2P(z,v)=(z_2,v_2)$ is given using group structure. $$z_2=\dfrac{(n^2-1)(4n^{10}-15n^8+26n^6-19n^4+9n^2-2)}{n(4n^{10}-17n^8+24n^6-18n^4+9n^2-3)}$$ $$v_2=-\dfrac{(n^4-1)(14-117n^2+498n^4-1367n^6+2661n^8-3768n^{10}+3863n^{12}-2802n^{14}+1353n^{16}-384n^{18}+48n^{20})}{(4n^{10}-17n^8+24n^6-18n^4+9n^2-3)^2n}$$ Similarly, we get $z_3=(n-1)(n+1)(64n^{30}-720n^{28}+3932n^{26}-13631n^{24}+32986n^{22}-58503n^{20}+78715n^{18}-82468n^{16}+68360n^{14}-45131n^{12}+23685n^{10}-9770n^8+3085n^6-709n^4+108n^2-8)\\ /(n(64n^{30}-816n^{28}+4604n^{26}-15521n^{24}+35500n^{22}-59320n^{20}+76107n^{18}-77744n^{16}+64806n^{14}-44595n^{12}+25324n^{10}-11727n^8+4307n^6-1193n^4+225n^2-22))$

Tomita
  • 3,209
  • 1
    Thanks. I've verified the edited version. I was hoping $z_2, z_3$ will be relatively simple. But they have deg-12, deg-32 numerators, and deg-11, deg-31 denominators, respectively. Oh, well, it would have been nice if the degrees were smaller. – Tito Piezas III Feb 15 '25 at 06:08
  • 1
    If it is not inconvenient, can you email me the data for $h<1000$, hopefully arranged by $a$? (Easier to spot patterns that way.) Maybe there is another family hidden in it. – Tito Piezas III Feb 15 '25 at 06:19
  • 1
    Ok, give me some time. – Tomita Feb 15 '25 at 06:20