Maximal displacement of a supercritical branching random walk in a time-inhomogeneous random environment (Q2274298): Difference between revisions

Let \(L:=(L_n)_{n\in\mathbb{N}}\) be a sequence of the distributions of point processes on the real line \(\mathbb{R}\). A time-inhomogeneous branching random walk in the environment \(L\) is defined as follows. It starts with one individual, the initial ancestor, located at the origin of \(\mathbb{R}\) at time \(0\). At time 1, the initial ancestor gives birth to children, with positions on \(\mathbb{R}\) given by a realization of a point process with distribution \(L_1\), and dies. All the children of the initial ancestor form the first generation. At time \(n\), each \((n-1)\)th generation individual gives birth to children, whose displacements w.r.t.\ the parent position are given by a realization of a point process with distribution \(L_n\), and for different parents these realizations are independent. The authors consider the situation in which \(L\) is a sequence of independent identically distributed random variables taking values in the space of the distributions of point processes. In this setting they investigate the asymptotic behavior of \(M_n\) the maximal position among all individuals at time \(n\) as \(n\to\infty\). The main result (Theorem 1.1) states that, conditionally on the environment \(L\), \(M_n\) behaves in probability as \(V_n-\varphi \log n+o(\log n)\), where \((V_n)\) is a random sequence and \(\varphi>0\) is a constant, both explicitly defined in the paper. The main technical tools exploited by the authors are the many-to-one lemma, a new ballot theorem for random walks in a random environment and some other nontrivial results.