Extinction time for the contact process on general graphs

TL;DR

This paper analyzes the extinction time of the contact process on general graphs, showing it grows super-exponentially with graph size and converges to an exponential distribution, extending previous bounded-degree results.

Contribution

It establishes super-exponential growth of extinction time on general graphs and proves convergence to exponential distribution, broadening earlier bounded-degree graph findings.

Findings

Extinction time grows faster than exponential in graph size for infection rates above critical.

Normalized extinction time converges to an exponential distribution as graph size increases.

Provides a simpler proof for exponential growth of extinction time in bounded-degree graphs.

Abstract

We consider the contact process on finite and connected graphs and study the behavior of the extinction time, that is, the amount of time that it takes for the infection to disappear in the process started from full occupancy. We prove, without any restriction on the graph $G$, that if the infection rate $\lambda$ is larger than the critical rate of the one-dimensional process, then the extinction time grows faster than $\exp\{|G|/(\log|G|)^\kappa\}$ for any constant $\kappa > 1$, where $|G|$ denotes the number of vertices of $G$. Also for general graphs, we show that the extinction time divided by its expectation converges in distribution, as the number of vertices tends to infinity, to the exponential distribution with parameter 1. These results complement earlier work of Mountford, Mourrat, Valesin and Yao, in which only graphs of bounded degrees were considered, and the extinction time was shown to grow exponentially in $n$; here we also provide a simpler proof of this fact.