2
$\begingroup$

I am asking this question mainly to probe the knowledge of people already familiar with this problem, otherwise I would advise caution to the unfamiliar trying to use computation, this can be really exhausting and frankly disappointing, hence my question, essentially made for people already acquainted with this problem. References, if any, would be greatly appreciated!

Let $R(t)$ be a rational function, like a polynomial $R(t)=t^2-t+1$ or a fraction of polynomials $R(t)=\frac{t^2+t+1}{t^3+2}$, etc.

Given $n$ (usually $n=2$), how do I know whether there exists $f$ such that

$f \circ f \circ f \cdots \circ f(t) = R(t)$ where the numbers of $\circ$ is $n-1$, and where $f$ is a power series with indices in $\mathbb{Z}$, i.e. $f$ has the form

$f(t)=\sum_{i=-\infty}^{+\infty}a_it^i$

and is there a nice algorithm to compute $f$?

To make things simpler, I am asking mainly about the algebraic existence (no topology), but it would be nice if the $f$'s could converge at least in some specific interval...

In any case, if $n=2$, can the rational functions obtained this way be characterized? If $n$ is any number $\geq 2$, can I get all the rational functions?

The reason I am asking this is that I've been through extremely tedious calculations to find solutions to recreational mathematics functional equations, usually like $f(f(x))=P(x)$, and while sometimes I was able to find such an $f$, sometimes I wasn't, for instance I think I spent way too much time trying to find $f$ satisfying $f(f(x))=x^2-x+1$. I don't know whether there exists a solution to this problem (it seems like it does not), the original problem was much simpler and did not require finding $f$.

$\endgroup$
10
  • $\begingroup$ We don't call $f\circ f\circ f$ a "power of" $f.$ $\endgroup$ Commented Nov 11 at 2:25
  • 1
    $\begingroup$ Title: you write "$n$-th power", but the Question appears to be about $n$-th iterated compositions. $\endgroup$ Commented Nov 11 at 2:25
  • $\begingroup$ Abstract power series cannot be composed, in general. It cannot be defined. You can compose these as abstract power series only when $f(0)=0.$ Otherwise, you end up with infinite sums for $f\circ f$ which might not converge. $\endgroup$ Commented Nov 11 at 2:30
  • 3
    $\begingroup$ People working in ergodic theory often refer to $f\circ f\circ f$ as a power of $f$. This might be due to papers such as digicoll.lib.berkeley.edu/record/113232/files/… $\endgroup$ Commented Nov 11 at 3:36
  • 1
    $\begingroup$ If $R$ is a polynomial, this is equivalent to recursively decomposing $R$. (For which, in the univariate case, polynomial time algorithms exist.) If the result is a composition of $k$ repetitions of the same polynomial, then $R$ is an iterated composition of that polynomial. (It is also an iterated composition of polynomials for each divisor of $k$ by grouping compositional factors.) Beware: decompositions are not unique. The interesting case seems to be $R$ not a polynomial. $\endgroup$ Commented Nov 11 at 3:57

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.