On the critical parameters of branching random walks (2507.14676v1)

Abstract: Given a discrete spatial structure $X$, we define continuous-time branching processes that model a population breeding and dying on $X$. These processes are usually called branching random walks. They are characterized by breeding rates $k_{xy}$ between sites (ruling the speed at which individuals at $x$ send children to $y$) and by the multiplicative speed parameter $\lambda$. Two critical parameters of interest are the global critical parameter $\lambda_w$, related to global survival, and the local critical parameter $\lambda_s$, related to survival in finite sets (with $\lambda_w\le\lambda_s)$. Local modifications of the rates of the process can affect the value of both parameters. Using results on the comparison of the probability of extinction of a single branching random walk on different sets, we are able to extend the comparison to extinction probabilities and critical parameters of couples of branching random walks, whose rates coincide outside a fixed set $A\subseteq X$. We say that two branching random walks are equivalent if their rates coincide everywhere but for a finite subset of $X$. Given an equivalence class of branching random walks, we prove that if one process has $\lambda^_w\neq\lambda^_s$, then $\lambda_w^*$ is the maximal possible value of this parameter in the class. We describe the possible situations for the critical parameters in these equivalence classes.