Critical branching processes in a random environment with immigration: the size of the only surviving family (Q2135145)

Consider a critical branching process \(\{Y_n, n \geq 0\}\) in an i.i.d. random environment in which one immigrant arrives at each generation. Specifically, at each generation \(n\), a random parameter \(p_n \in (0,1)\) is sampled at random. Then, writing \((\xi_{n+1,j})\) i.i.d. random variables with geometric distribution of parameter \(p_n\), the process at time \(n+1\) is defined as \(Y_{n+1} = 1 + \sum_{j=1}^{Y_n} \xi_{n+1,j}\). Denote by \(S_n = \sum_{j=1}^n \log \frac{p_j}{1-p_j}\) the associated random walk to the process \(Y\). The branching process in random environment \(\{Y_n\}\) is assumed to be critical, i.e., \(\mathbb{E}[\log \frac{p_n}{1-p_n}] =0\), with \(\mathbb{E}[\frac{1}{p_n} + \frac{1}{1-p_n}] < \infty\). For \(0 \leq i \leq n\), let \(\mathcal{A}_i(n)\) be the event that all individuals alive at time \(n\) are offspring of the immigrant which joined the population at time \(i\). This article describes the asymptotic behaviour of \(Y_n\) conditioned on \(\mathcal{A}_i(n)\) as \(n \to \infty\), depending on the asymptotic behaviour of \(i\). Whether \(i\) remains fixed, \(n-i\) remains fixed or \(i\) and \(n-i\) both grow to \(\infty\), \(Y_n\) is showed to be of the same order of magnitude as \(e^{S_n -S_i}\), conditionally on \(\mathcal{A}_i(n)\).