Branching property for a Poisson-type death process (Q1586592)

Let \(\xi_t\), \(t\geq 0\), be a time-homogeneous Markov process with state space \(\{0,1,2,\dots\}\). Denote the transition probabilities of \(\xi_t\) by \[ P_{ij}(t)={\mathbf P}(\xi_t=j\mid \xi_0=i), \quad i,j=0,1,2,\dots . \] Assume that \[ P_{i,i-1}(t)=\mu_i t+o(t),\quad P_{ii}(t)=1-\mu_i t+o(t), \] where \(\mu_0=0\), \(\mu_i\geq 0\), \(i=0,1,\dots\). Introduce the generating function of the transition probabilities \[ F_i(t;s)=\sum_{j=0}^\infty P_{ij}(t)s^j, \quad i,j=0,1,2,\dots , \quad |s|\leq 1. \] In the case of a linear death process, i.e.\ \(\mu_i=i\mu\) (\(\mu>0\)), the branching property takes place: \[ F_i(t;s)=F_1^i(t;s)=(1-e^{-\mu t}+se^{-\mu t})^i, \] see \textit{B. A. Sevastyanov} [``Branching processes'' (1971; Zbl 0238.60001)]. In the present paper the author derives the explicit formulae for \(F_i(t;s)\) in the case of the Poisson-type death process with \(\mu_0=0\), \(\mu_i=\mu\), \(i=1,2,\dots\), and the square-type death process with \(\mu_i=i(i-1)\mu\).