Borel Vizing's Theorem for Graphs of Subexponential Growth
Anton Bernshteyn, Abhishek Dhawan
TL;DR
This paper proves a Borel analogue of Vizing's theorem for graphs of subexponential growth: any Borel graph $G$ with finite maximum degree $\Delta$ and subexponential growth satisfies $\chi'_{\mathsf{B}}(G) \le \Delta + 1$. The authors derive a stronger finite-result—a deterministic $\mathsf{LOCAL}$ algorithm colors an $n$-vertex graph of subexponential growth with $\Delta+1$ colors in $O(\log^* n)$ rounds, with constants depending on the growth bound—by leveraging Christiansen's small augmenting subgraphs. They then transfer this finite-case outcome to the Borel setting via Bernshteyn's transfer principle, establishing the global bound for Borel graphs. The work combines descriptive set theoretic and distributed computing techniques, yielding a robust bridge between finite algorithmic guarantees and Borel graph colorings, and extends the scope of Vizing-type results to broader growth regimes. This advances the understanding of when combinatorial colorings admit Borel realizations and demonstrates the practical impact of distributed-algorithm tools in descriptive combinatorics.
Abstract
We show that every Borel graph $G$ of subexponential growth has a Borel proper edge-coloring with $Δ(G) + 1$ colors. We deduce this from a stronger result, namely that an $n$-vertex (finite) graph $G$ of subexponential growth can be properly edge-colored using $Δ(G) + 1$ colors by an $O(\log^\ast n)$-round deterministic distributed algorithm in the $\mathsf{LOCAL}$ model, where the implied constants in the $O(\cdot)$ notation are determined by a bound on the growth rate of $G$.
