2
$\begingroup$

Does the Diophantine equation $y^2=x^2-x^3+x^4$ have solutions other than $x=1,y=1$? Interestingly, the Diophantine equation $y^2=x^2-x^3+x^5$ has such solutions: $x=3,y=15$, $x=5,y=55$, $x=56,y=23464$. However I don't know if there are any other solutions. The solution $x=3,y=15$ is related to the Mordell's equations $y^3=x^2+2$ and $y^3=x^2+18$ which have well known sole solutions $x=5,y=3$ and $x=3,y=3$. Indeed it follows from $3^3=5^2+2$ and $3^3=3^2+18$ that $3^5-(15)^2-(3^3-3^2)=0$.

$\endgroup$
  • 2
    $\begingroup$ The solutions to $y^{2} = x^{2} - x^{3} + x^{5}$ all come from the integral points on $E : y^{2} = x^{3} - x + 1$. Magma indicates that the only solutions are $x = -1$, $x = 0$, $x = 1$, $x = 3$, $x = 5$ and $x = 56$. $\endgroup$ – Jeremy Rouse Jul 23 '14 at 12:18
12
$\begingroup$

The only integer solutions $(x,y)$ are $(0,0)$ and $(1,\pm 1)$.

Your equation is $y^2=x^2(1-x+x^2)$. This can only have integer solutions if $p(x)=1-x+x^2$ is a square or $y=0$. For $x>1$ we have $(x-1)^2<p(x)<x^2$, so $p(x)$ is not a square. For $x<0$ we have $x^2<p(x)<(x-1)^2$, so $p(x)$ is not a square. The only possibilities are $x=0$ and $x=1$, and they indeed give solutions.

$\endgroup$
  • 2
    $\begingroup$ Alternately, for $x\neq 0$, $z:=y/x$ is an integer, and the equation factors as $(2z-2x+1)(2z+2x-1)=3$. It is easy to finish from here. $\endgroup$ – GH from MO Jul 23 '14 at 12:25
  • 1
    $\begingroup$ Or $z^{2} = (x+\omega)(x+\omega^{2}),$ where $\omega = e^{\frac{2 \pi i }{3}}$ (and $z$ must be prime to $3$). This leads (if $x+\omega$ is not a unit) to $(x+ \omega) = (a +b\omega)^{2}$ for integers $a$ and $b.$ Hence $1 = 2ab-b^{2},$ so that either $b= a = 1$ and $x =0$ or $b = a = -1, x =0.$ The possibility $x + \omega = -\omega^{2}$ leads to $x=1$. $\endgroup$ – Geoff Robinson Jul 23 '14 at 12:46
  • $\begingroup$ I think Joonas solution is not complete. For example we have to prove that the system $x^2=a^3,p(x)=a$ does not have integer solutions for some integer $a$. $\endgroup$ – Zurab Silagadze Jul 23 '14 at 12:59
  • 1
    $\begingroup$ @ZurabSilagadze: Joonas's solution is fine. If $y^2=x^2(1-x+x^2)$ holds for $y\neq 0$, then $1-x+x^2$ is a rational square and also an integer, hence an integral square. On the other hand, he shows that unless $x=0$ or $x=1$ holds, $1-x+x^2$ lies strictly between two neighboring integral squares, so it is not an integral square. I outlined an other approach in my comment, which avoids inequalities. It is based on the fact that $3$ cannot be factored in many ways over the integers. $\endgroup$ – GH from MO Jul 23 '14 at 13:06
  • 2
    $\begingroup$ @GHfromMO: Thanks, now I see the point. I have accepted Joonas's solution as the correct one. You solution is cute. $\endgroup$ – Zurab Silagadze Jul 24 '14 at 8:46

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.