The Markov branching-castastrophe process (Q1096562)

A continuous time Markov branching process is modified by independent random decrements in the population size occurring at a rate proportional to the current population size. If a decrement variable is larger than the current population size then the population is extinguished. The special case of a birth-death-immigration process with three specific forms of decrement distribution has been considered by \textit{P. J. Brockwell}, Adv. Appl. Probab. 17, 42-52 (1985; Zbl 0551.92013); see also \textit{P. J. Brockwell}, \textit{J. Gani} and \textit{S. I. Resnick}, ibid. 14, 709-731 (1982; Zbl 0496.92007). In the present paper three problems are considered: (i) whether extinction is almost sure and, if not, how the extinction probability \(q_ i\) behaves asymptotically as a function of the initial population size i; (ii) the behaviour of the mean time to extinction as a function of i in the case of almost sure extinction; (iii) the asymptotic growth of the population in the event of non-extinction. Use is made of the fact that the embedded jump process is a random walk with absorbing barrier at 0. Several questions are posed but left open. The results are made more explicit in the special case of a geometric decrement distribution.