English

Percolation through Isoperimetry

Combinatorics 2024-01-19 v3 Probability

Abstract

We provide a sufficient condition on the isoperimetric properties of a regular graph GG of growing degree dd, under which the random subgraph GpG_p typically undergoes a phase transition around p=1dp=\frac{1}{d} which resembles the emergence of a giant component in the binomial random graph model G(n,p)G(n,p). We further show that this condition is tight. More precisely, let d=ω(1)d=\omega(1), let ϵ>0\epsilon>0 be a small enough constant, and let pd=1+ϵp \cdot d=1+\epsilon. We show that if CC is sufficiently large and GG is a dd-regular nn-vertex graph where every subset SV(G)S\subseteq V(G) of order at most n2\frac{n}{2} has edge-boundary of size at least CSC|S|, then GpG_p typically has a unique linear sized component, whose order is asymptotically y(ϵ)ny(\epsilon)n, where y(ϵ)y(\epsilon) is the survival probability of a Galton-Watson tree with offspring distribution Po(1+ϵ)(1+\epsilon). We further give examples to show that this result is tight both in terms of its dependence on CC, and with respect to the order of the second-largest component. We also consider a more general setting, where we only control the expansion of sets up to size kk. In this case, we show that if GG is such that every subset SV(G)S\subseteq V(G) of order at most kk has edge-boundary of size at least dSd|S| and pp is such that pd1+ϵp\cdot d \geq 1 + \epsilon, then GpG_p typically contains a component of order Ω(k)\Omega(k).

Keywords

Cite

@article{arxiv.2308.10267,
  title  = {Percolation through Isoperimetry},
  author = {Sahar Diskin and Joshua Erde and Mihyun Kang and Michael Krivelevich},
  journal= {arXiv preprint arXiv:2308.10267},
  year   = {2024}
}
R2 v1 2026-06-28T11:59:46.499Z