2
$\begingroup$

The normalized Fermat curve is $X^n+Y^n=1$. We have of course infinite possibilities for parametrisation, but the periodicity is a special characteristic here.

E.g. $\cos^2x+\sin^2x=1$ has periodicity, but not $(\frac{1-x^2}{1+x^2})^2+(\frac{2x}{1+x^2})^2=1$.

Here I am interested in periodic functions for $n:=3$. Be $x\in \mathbb{C}$.

Be $f^\prime(x)=-\frac{g^2}{f}(x)$ (note: pole $x_0$ for $f(x_0)=0$) , $g^\prime(x)=f(x)$, $f(0)=1$ and $g(0)=0$.

=> $f^3(x)+g^3(x)=1$

Be $\eta$ and $\mu$ such that $f(\eta)=0,\, g(\eta)=1$ and $f(\mu)=g(\mu)=\frac{1}{\sqrt[3]{2}}$.

$X^{-1}$ here is used as inverse function for $X$ within a valid value range. After a few transformations one get e.g.

$(\frac{g}{f})^{-1}(x)=\int\limits_0^x \frac{dt}{1+t^3}$,

$g^{-1}(x)=\int\limits_0^x \frac{dt}{\sqrt[3]{1+t^3}}$,

$f^{-1}(x)=\int\limits_x^1 \frac{tdt}{\sqrt[3]{1+t^3}^2}$,

$\mu=(\frac{g}{f})^{-1}(1)=\frac{\pi}{3\sqrt{3}}+\frac{1}{3}\ln 2$,

$\eta=\frac{\Gamma(\frac{1}{3})\Gamma(\frac{2}{3})}{\Gamma(1)}=\frac{2\pi}{3\sqrt{3}}=2\mu-\frac{1}{3}\ln 4$.

Be $\lambda:=e^{i\frac{2\pi}{3}}$.

Note: With the Taylor expansion around 0 for $f(x)$ and $g(x)$ one get $f(\lambda x)=f(x)$ and $g(\lambda x)=\lambda g(x)$.

The addition theorems for $g$ and $f$.

With $(\frac{g}{f})^{-1}(x)=\frac{1}{3}(\ln(1+x)+\lambda^2\ln(1+\lambda x)+\lambda\ln(1+\lambda^2 x))$ follows $(\frac{g}{f})^{-1}(x)+(\frac{g}{f})^{-1}(y)+\frac{1}{3}\ln(1-\frac{3xy}{(1+x)(1+y)})=(\frac{g}{f})^{-1}(\frac{x+y-xy}{1-xy})$

and with $x\to\frac{g}{f}(x)$, $y\to\frac{g}{f}(y)$ and $f^3(z)+g^3(z)=1$ one get

$f(x+y+\frac{1}{3}\ln(1-\frac{3g(x)g(y)}{(f(x)+g(x))(f(y)+g(y))}))=$ $\frac{f(x)f(y)-g(x)g(y)}{\sqrt[3]{(f(x)f(y)-g(x)g(y))^3+(g(x)f(y)+f(x)g(y)-g(x)g(y))^3}}$ and

$g(x+y+\frac{1}{3}\ln(1-\frac{3g(x)g(y)}{(f(x)+g(x))(f(y)+g(y))}))=$ $\frac{g(x)f(y)+f(x)g(y)-g(x)g(y)}{\sqrt[3]{(f(x)f(y)-g(x)g(y))^3+(g(x)f(y)+f(x)g(y)-g(x)g(y))^3}}$

Be $\phi:=\eta+\mu-\frac{1}{3}\ln(-2)=\frac{\pi}{\sqrt{3}}-\frac{\ln(-1)}{3}$.

With $x:=\eta$ and $y:=\mu$ one get $g(\phi)=0$ and $f(\phi)=1$.

General: $g(a)=0$ => $g(x+a)=g(x)$ , here: $\Re(a)\ge 0$.

$x:=\lambda a$ and $y:=\lambda^2 a$: $0=g(0)=g(-a)$ , one get $g(\pm a)=0$

QUESTION: How can we interpret $\ln(-1)$ here ?

  • Does it mean that $g(x+k\cdot\frac{\pi}{\sqrt{3}}+l\cdot\frac{i\pi}{3})=g(x)$ ? $k,l\in\mathbb{Z}$, maybe only $k\in\mathbb{N_0}$

  • Does $l$ depend on $k$ or are they independent of each other ?

A supplementary question: Has $\frac{g(\phi x)}{\phi x}$ anything to do with $\prod\limits_{k=1}^\infty (1-(\frac{x}{k})^3)$ ?

EDIT:

Because of the comments below, the double periodicity here has the form

$g(x+k(\frac{\pi}{\sqrt{3}}+i(2l+1)\frac{\pi}{3}))=g(x)$

with $k,l\in\mathbb{Z}$.

$\endgroup$
13
  • $\begingroup$ You have partly rediscovered the historical background for theory of elliptic functions (and elliptic integrals) in the case of the elliptic curve $x^3+y^3=1$... Enchanting stuff, but perhaps better suited for MathSE. A slightly different example than the Fermat curve is written up in some course notes at math.umn.edu/~garrett/m/complex/10_elliptic.pdf $\endgroup$ Jun 19, 2016 at 21:49
  • $\begingroup$ Thanks for the hint - but is this a reason to downvote ? I couldn' get the pdf, Proxy error. I know elliptic functions, but have no answer to my question. I am looking for someone, who can handle such problems. $\endgroup$
    – user90369
    Jun 20, 2016 at 5:49
  • $\begingroup$ DIdn't downvote... $\endgroup$ Jun 20, 2016 at 11:58
  • $\begingroup$ O.k. sorry Paul Garrett. - And to the unknown person/moderator: If somebody downvotes, he/she should say why (arguments!). $\endgroup$
    – user90369
    Jun 20, 2016 at 21:50
  • 1
    $\begingroup$ Yes, ideally, one might hope for some feedback about downvotes. But, yes, perhaps disappointingly, it should also (in my opinion, and in some others') be ok to downvote anonymously. Sure, this could be abused... But, srsly, all this voting stuff is just a popularity contest, anyway. So don't be put off by it... if this venue provides you worthwhile information. Things here are very rarely as idiotic as many other places on the introweblogs. $\endgroup$ Jun 20, 2016 at 22:18

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.