3
$\begingroup$

I 'm searching about the solvability of the functional equation $f(f(x))=g(x)$. I have three questions about it:

  1. Let's be $g$ an arbitrary function and the functional equation $f(f(x))=g(x)$. Are there any specific criteria to ensure us that there exist such function $f$?
  2. Let's be $g\in C^0$ and the functional equation $f(f(x))=g(x)$. Are there any specific criteria to ensure us that there exist such function $f$ and $f\in C^0$?
  3. Let's be $g\in C^1$ and the functional equation $f(f(x))=g(x)$. Are there any specific criteria to ensure us that there exist such function $f$ and $f\in C^1$?

Thanks in advance!

P.S.: I read this and this, but there are a little bit different questions.

$\endgroup$
4
  • $\begingroup$ The question may be too broad. Do you have any specific examples in mind? $\endgroup$ Commented Oct 3, 2019 at 8:26
  • 3
    $\begingroup$ The answer in 1 would be Ulm invariants. See my answer mathoverflow.net/a/17869/454 $\endgroup$ Commented Oct 3, 2019 at 9:20
  • $\begingroup$ @Bullet51 , I hope that there will be something like the discriminant on the quadratic equations. $\endgroup$ Commented Oct 6, 2019 at 11:06
  • $\begingroup$ @GeraldEdgar , thanks for your answer!!! $\endgroup$ Commented Oct 6, 2019 at 11:06

0

You must log in to answer this question.

Browse other questions tagged .