On the Graphical $r$-Stirling Numbers of the First Kind for Specific Graph Families
Daniel Yaqubi, Madjid Mirzavaziri
TL;DR
This work introduces graphical $r$-Stirling numbers of the first kind ${G \brack k}_r$, unifying restricted permutation cycle structures with graph-based cycle partitions. It develops explicit closed forms, recurrences, and $r$-cycle polynomials $\mathcal{C}_r(G,x)$ for families such as $P_n$, $C_n$, $S_n$, $W_n$, and $F_n$, and derives mean and variance formulas for the cycle-count distributions, including asymptotic normality in several regimes. The analysis leverages bridge-decomposition, generating polynomials, and Fibonacci/Lucas polynomials, connecting algebraic properties (real-rootedness, log-concavity) with probabilistic limit laws. The results provide a robust framework for restricted cycle decompositions in graphs and pave the way for further exploration of asymptotics, unimodality, and real-rootedness across broader graph classes.
Abstract
This paper investigates the \textbf{graphical $r$-Stirling numbers of the first kind}, denoted by $\str{G}{k}$, which enumerate partitions of a vertex set $V(G)$ into $k$ disjoint cycles such that $r$ specified vertices occupy distinct blocks. We establish closed-form expressions and recursive identities for fundamental graph families, including \textbf{Path} ($P_n$), \textbf{Cycle} ($C_n$), \textbf{Star} ($S_n$), \textbf{Wheel} ($W_n$), and \textbf{Fan} ($F_n$) graphs. A primary focus of this study is the \textbf{statistical characterization} of the cycle distribution. We derive explicit formulas for the \textbf{mean} and \textbf{variance} of these numbers, extracted from the structural properties of the $r$-cycle polynomials. These results provide a rigorous measure of the average cycle density and variability across different graph topologies, bridging the gap between algebraic combinatorics and the structural analysis of restricted permutations.
