References & Citations
Mathematics > Probability
Title: Bernoulli percolation on the Random Geometric Graph
(Submitted on 22 May 2022)
Abstract: Given $\lambda > 0$, $p\in [0,1]$ and a Poisson Point Process $\mathrm{Po}(\lambda)$ in $\mathbb R^2$ with intensity $\lambda$, we consider the random graph $G = G(\lambda, p)$ with vertex set $\mathrm{Po}(\lambda)$ in which every pair of vertices at distance at most 1 forms an edge with probability $p$, independently of other pairs. Our study is motivated by the question of locality for bond percolation on the standard random geometric graph $G(\lambda,1)$. We show that for a large class of graphs converging locally to $G(\lambda,1)$ in a suitable sense, the corresponding critical percolation thresholds converge to the one of $G(\lambda,1)$. The proof is based on a finite volume criterion. In this direction, we strengthen recent results of Penrose by showing that the size of the largest component rescaled by $n$ converges almost surely to a constant, and providing sharp bounds for the size of the second-largest component.
Link back to: arXiv, form interface, contact.