7
$\begingroup$

Let $S$ be a finite set and $f: S \to S$ be a function. Let $k = |f(S)|$ and let $\alpha$ be the partition of $S$ into $f$-fibers, i.e. $\alpha = \{ \alpha_t \}_{t \in f(S)}$ where $\alpha_t = f^{-1}(\{t \})$. Then $\sum_{t \in f(S)} |\alpha_t|$ is a partition of $|S|$ with $k$ nonempty parts.

I would really like to have a nice expression, or indeed bijection, for the number of pairs $g, h: S \to S$ such that $f = g \circ h$. These can be enumerated in a naive way with Bell and Stirling numbers of the second kind, by considering that the $h$-fibers need to refine $\alpha$ and $g(S)$ needs to contain $f(S)$ (and the $g$-fibers need to partition $h(S)$ in the right way), but the expression is not enlightening at all. If there's some nice identity or known bijective correspondence to which someone could point me, I'd be very grateful.

$\endgroup$
2
  • 4
    $\begingroup$ Suppose $f$ is a constant function, i.e., $\alpha=\{S\}$. Let $n=\#S$. Then the number of such factorizations is $\sum_{i=0}^{n} n^{n-i} \cdot (n)_i \cdot S(n,i)$ where $S(n,i)$ is the Stirling number of the 2nd kind and $(n)_i := n(n-1)\cdots(n-i+1)$. Already in this special case it does not seem like we can do better than a summation. For example, with $n=8$ this evaluates to $28186856832 = 2^7\times 3\times 73403273$, which has a huge prime factor (so there cannot be a product formula). $\endgroup$ Commented Nov 9, 2021 at 22:46
  • 1
    $\begingroup$ @SamHopkins thanks for doing that calculation! Indeed, the constant maps are maximal. More generally, if $f_1, f_2$ are such that the partition into $f_1$-fibers refines the partition into $f_2$-fibers, then there are at least as many factorizations of $f_2$ as of $f_1$. (Any $h$ that works for $f_2$ will work for $f_1$, and once you've picked $h$, say $|h(S)| = m$, you have $n^{n-m}$ choices for $g$, whether factoring $f_1$ or $f_2$.) $\endgroup$
    – Sophie M
    Commented Nov 10, 2021 at 2:07

0

You must log in to answer this question.