Random 2-cell embeddings of multistars
Jesse Campion Loth, Kevin Halasz, Tomáš Masařík, Bojan Mohar, Robert Šámal
TL;DR
This work analyzes the average number of faces $\mathbb{E}[F_G]$ in uniform random orientable embeddings of a graph $G$, equating to the average genus. It uses local-rotation representations and Stanley's generating functions to obtain an exact expression for the dipole case and a tight, near-dipole bound for multistars, depending only on the partition type and a simple harmonic-mean quantity $\Delta_{n'}$. These results are extended to general graphs via a vertex-adding framework that yields practical upper bounds such as $\mathbb{E}[F] = O(n)$ in many families and a bound $\mathbb{E}[F] < n(\log d+5/3)$ for graphs of maximum degree $d$, along with $d$-degenerate bounds. The paper also posits a conjecture that $\mathbb{E}[F_G] = O(n)$ for all simple graphs, supported by linear-growth examples and degree-based analyses. Overall, it provides precise tools to bound topological complexity of random graph embeddings and connects dipole/multistar structures to broader graph classes.
Abstract
Random 2-cell embeddings of a given graph $G$ are obtained by choosing a random local rotation around every vertex. We analyze the expected number of faces, $\mathbb{E}[F_G]$, of such an embedding which is equivalent to studying its average genus. So far, tight results are known for two families called monopoles and dipoles. We extend the dipole result to a more general family called multistars, i.e., loopless multigraphs in which there is a vertex incident with all the edges. In particular, we show that the expected number of faces of every multistar with $n$ nonleaf edges lies in an interval of length $2/(n + 1)$ centered at the expected number of faces of an $n$-edge dipole. This allows us to derive bounds on $\mathbb{E}[F_G]$ for any given graph $G$ in terms of vertex degrees. We conjecture that $\mathbb{E}[F_G ] \le O(n)$ for any simple $n$-vertex graph $G$.
