0
$\begingroup$

How to transform the projective equation $$X^3+Y^3+Z^3-3\mu XYZ = 0$$ into the standard form of elliptic curve $$ y^2=x^3+px+q? $$

$\endgroup$
1
$\begingroup$

Here's a really terrible answer: ask a computer. I asked a computer ($\textsf{Magma}$, to be specific), and here's what it told me:

> QQu<u> := RationalFunctionField(Rationals());
> PP<X,Y,Z> := ProjectiveSpace(QQu,2);
> R<T0,T1,T2> := PolynomialRing(QQu,3);
> f := T0^3 + T1^3 + T2^3 - 3*u*T0*T1*T2;
> C := Curve(PP,f);
> E, phi := EllipticCurve(C);
> E;
Elliptic Curve defined by y^2 + 3*u/(u^3 - 1)*x*y - 9/(u^6 - 2*u^3 + 1)*y = x^3 - 9*u^2/(u^6 - 2*u^3 + 1)*x^2 + 27*u/(u^9 - 3*u^6 + 3*u^3 - 1)*x - 27/(u^12 - 
    4*u^9 + 6*u^6 - 4*u^3 + 1) over Univariate rational function field over Rational Field
> phi;
Mapping from: CrvPln: C to CrvEll: E
Composition of Mapping from: CrvPln: C to Curve over Univariate rational function field over Rational Field defined by
X^3 - 3*u^2/(u^3 - 1)*X^2*Z + 3*u/(u^3 - 1)*X*Y*Z - 3/(u^3 - 1)*Y^2*Z + 3*u/(u^3 - 1)*X*Z^2 - 3/(u^3 - 1)*Y*Z^2 - 1/(u^3 - 1)*Z^3
with equations : 
Y
-X
X + u*Y + Z and
Mapping from: Curve over Univariate rational function field over Rational Field defined by
X^3 - 3*u^2/(u^3 - 1)*X^2*Z + 3*u/(u^3 - 1)*X*Y*Z - 3/(u^3 - 1)*Y^2*Z + 3*u/(u^3 - 1)*X*Z^2 - 3/(u^3 - 1)*Y*Z^2 - 1/(u^3 - 1)*Z^3 to CrvEll: E
with equations : 
3/(u^3 - 1)*X
-9/(u^6 - 2*u^3 + 1)*Y
Z
and inverse
-3/(u^3 - 1)*$.1
$.2
-9/(u^6 - 2*u^3 + 1)*$.3
> Ew, psi, psi_inv := WeierstrassModel(E);
> Ew;
Elliptic Curve defined by y^2 = x^3 + (-2187*u^4 - 17496*u)/(u^12 - 4*u^9 + 6*u^6 - 4*u^3 + 1)*x + (39366*u^6 - 787320*u^3 - 314928)/(u^18 - 6*u^15 + 15*u^12 
    - 20*u^9 + 15*u^6 - 6*u^3 + 1) over Univariate rational function field over Rational Field
> psi;
Elliptic curve isomorphism from: CrvEll: E to CrvEll: Ew
Taking (x : y : 1) to (36*x - 81*u^2/(u^6 - 2*u^3 + 1) : 216*y + 324*u/(u^3 - 1)*x - 972/(u^6 - 2*u^3 + 1) : 1)
> psi_inv;
Elliptic curve isomorphism from: CrvEll: Ew to CrvEll: E
Taking (x : y : 1) to (1/36*x + 9/4*u^2/(u^6 - 2*u^3 + 1) : 1/216*y - 1/24*u/(u^3 - 1)*x + (9/8*u^3 - 9/2)/(u^9 - 3*u^6 + 3*u^3 - 1) : 1)

(Here $u$ is $\mu$.) Let me try to translate this into English. $\textsf{Magma}$ first gave me a curve $E$ in long Weierstrass form: \begin{align*} y^2 + \frac{3u}{u^3 - 1} xy &- \frac{9}{u^6 - 2u^3 + 1} y =\\ &x^3 - 9 \frac{u^2}{u^6 - 2u^3 + 1} x^2 + \frac{27 u}{u^9 - 3u^6 + 3u^3 - 1} x - \frac{27}{u^{12} - 4u^9 + 6u^6 - 4u^3 + 1} \end{align*} It also gave me an isomorphism $\phi: C \overset{\sim}{\to} E$ as a composition of maps.

Next, I asked for a curve in short Weierstrass form, and $\textsf{Magma}$ gave me the curve $E_w$: \begin{align*} y^2 &= x^3 + \frac{-2187 u^4 - 17496 u}{u^{12} - 4 u^9 + 6 u^6 - 4 u^3 + 1}x + \frac{39366 u^6 - 787320 u^3 - 314928}{u^{18} - 6 u^{15} + 15 u^{12} - 20 u^9 + 15 u^6 - 6 u^3 + 1} \end{align*} as well as an isomorphism $\psi: E \overset{\sim}{\to} E_w$ and its inverse. This looks pretty horrible, and there may very well be a much nicer answer. I also tried asking for a minimal model, but it would take me a while to get it to fit it on the screen. One last word of warning: $\textsf{Magma}$ required that the base ring be a field, so $\mu$ is not allowed to be $0$ (despite the fact that $X^3 + Y^3 + Z^3 = 0$ defines a perfectly good elliptic curve), although the equations might still work out.

$\endgroup$

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.