Self-similar growth fragmentations as scaling limits of Markov branching processes (Q2181609)

The purpose of the paper is to study Markov branching systems in discrete time and space. The processes considered are particle systems that at each time step are a collection of particles of certain sizes. There is a Markov dynamic where at each step, every particle is either replaced by a bigger one (growth) or by two smaller particles conserving the total size of the parent particle (conservative fragmentation). The transition happens with probability \(p_{n,m}\) if \(n\) is the size of the parent particle and \(m\) denotes the size of the largest of the (one or two) resulting particles. Let \(X\) denote a Markov chain on \(\mathbb{N}\) with transition probabilities \(p_{n,m}\). The basic assumption is the convergence in distribution of the rescaled (in time and space) Markov chain \(X\) in the Skorokhod space of right-continuous functions with left limits towards a positive strong Markov process \(Y=(Y(t))_{t \geq 0}\), continuously absorbed at \(0\) and with a certain self-similarity property. Under an additional assumption, the author proves the convergence in the Skorokhod space of the Markov branching system in distribution to the self-similar growth fragmentation driven by \(Y\). He also derives a scaling limit for the genealogical embedding considered as a compact tree.