A solution of the functional equation $$f\left(\frac{1}{x}\right)=\sqrt{x}f(x),\quad x\gt 0$$ is $$f(x)=\sum_{n=-\infty}^\infty e^{-n^2\pi x}.$$ Another solution should be a certain quadratic function, though I encountered this problem a long time ago and forgot that function (together with the approach).

So there is one or (maybe) two or more solutions of the above functional equation. But how can I prove, given a set of solutions, that these are all the solutions of the equation?

By the way, does anyone know that quadratic function (if it exists at all)?

  • 1
    $\begingroup$ $f(x) =x^{-1/4}$ satisfies the functional equation. Also any constant multiple of a function that satisfies the equation also satisfies it. $\endgroup$ – eyeballfrog Mar 16 at 17:29
  • 2
    $\begingroup$ If $f$ and $g$ are positive functions satisfying the equation, for any $t \in \mathbb{R}$, $f^tg^{1-t}$ satisfies the equation too. $\endgroup$ – Mindlack Mar 16 at 17:52
  • $\begingroup$ @Yves Daoust $\sum_{n=-\infty}^\infty e^{-n^2\pi x}=\vartheta _{3}(0,e^{-\pi x})$ where $\vartheta$ is the Jacobi theta function. It converges for $x\gt 0$. mathworld.wolfram.com/JacobiThetaFunctions.html $\endgroup$ – Poder Rac Mar 16 at 17:54
  • $\begingroup$ If by a quadratic function you mean $f(x)=ax^2+bx+c$ then I believe no such solutions exist $\endgroup$ – მამუკა ჯიბლაძე Mar 16 at 17:57
  • $\begingroup$ @მამუკა ჯიბლაძე I may have erred, then. $\endgroup$ – Poder Rac Mar 16 at 17:58

These are not all solutions to the equation. We can simply define

$$f(x)=\begin{cases}g(x),&x\ge1\\g(1/x)/\sqrt x,&x<1\end{cases}$$

for any function $g$.

| cite | improve this answer | |

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.