Finding long cycles in a percolated expander graphs

TL;DR

This paper proves that in certain vertex-expander graphs, a percolated subgraph with a specific retention probability almost surely contains a long cycle proportional to the graph's parameters, extending previous results.

Contribution

It extends prior work by demonstrating the existence of long cycles in percolated expander graphs under a stronger conjectured condition.

Findings

Long cycles of length proportional to kd exist in percolated graphs

Probability of existence approaches 1 exponentially fast as k increases

Results hold for graphs with strong vertex expansion properties

Abstract

Given a graph $G$, the percolated graph $G_p$ has each edge independently retained with probability $p$. Collares, Diskin, Erde, and Krivelevich initiated the study of large structures in percolated single-scale vertex expander graphs, wherein every set of exactly $k$ vertices of $G$ has at least $dk$ neighbours before percolation. We extend their result to a conjectured stronger form, proving that if $p = (1+\varepsilon)/d$ and $G$ is a graph on at least $k$ vertices which expands as above, then $G_p$ contains a cycle of length $\Omega_\varepsilon(kd)$ with probability at least $1-\exp(-\Omega_\varepsilon(k/d))$ as $k\rightarrow\infty$.