English

Bounds for the independence and chromatic numbers of locally sparse graphs

Combinatorics 2025-07-22 v3

Abstract

In this note we consider a more general version of local sparsity introduced recently by Anderson, Kuchukova, and the author. In particular, we say a graph G=(V,E)G = (V, E) is (k,r)(k, r)-locally-sparse if for each vertex vV(G)v \in V(G), the subgraph induced by its neighborhood contains at most kk cliques of size rr. For r3r \geq 3 and ϵ[0,1]\epsilon \in [0, 1], we show that an nn-vertex (Δϵr,r)(\Delta^{\epsilon r}, r)-locally-sparse graph GG of maximum degree Δ\Delta satisfies α(G)(1o(1))nηΔ\alpha(G) \geq (1-o(1))\dfrac{n}{\eta\Delta} and χ(G)Θ(ηΔ)\chi(G) \leq \Theta\left(\eta\Delta\right), where η:=ϵ+rloglogΔlogΔ\eta := \epsilon + \dfrac{r\log\log \Delta}{\log \Delta}. For ϵ\epsilon not too large, the hidden constant in the Θ()\Theta(\cdot) can be taken to be 1+o(1)1+o(1). Setting ϵ=0\epsilon = 0, we recover classical results on Kr+1K_{r+1}-free graphs due to Shearer and Johansson, which were more recently improved by Davies, Kang, Pirot, and Sereni. We prove a stronger result on the independence number in terms of the occupancy fraction in the hard-core model, and establish a local version of the coloring result in the more general setting of correspondence coloring.

Keywords

Cite

@article{arxiv.2403.03054,
  title  = {Bounds for the independence and chromatic numbers of locally sparse graphs},
  author = {Abhishek Dhawan},
  journal= {arXiv preprint arXiv:2403.03054},
  year   = {2025}
}

Comments

23 pages