On the inequalities for moments of branching random processes (Q6154858)

Consider a Galton-Watson process with immigration, defined recursively for all \(n \in \mathbb N\) by \[ X_n = \eta_n + \sum_{i=1}^{X_{n-1}} \xi_{n,i}, \] where \(\{X_0\}, \{\eta_j, j \in \mathbb N\}\) and \(\{\xi_{i,j}, (i,j) \in \mathbb{N}^2\}\) are three independent families of i.i.d. integer-valued random variables on \(Z_+\). The present articles gives explicit bounds on the moments \(E X_n^p\) and \(E|X_n-E(X_n)|^p\) for all \(p \in (0,\infty)\), depending only on the moments of the laws of \(X_0\), \(\eta_1\) and \(\xi_{1,1}\). The proofs are mainly based on the adaptation of arguments used to study the moments of sums of independent random variables, such as von Bahr-Essen and Marcinkiewicz-Zygmund inequalities, to sums of a random number of independent random variables.