On the geometric growth in controlled branching processes with random control function

From MaRDI portal





Let \(\{X_{nj}: n=0,1,\dots, j=1,2,\dots\}\) and \(\{\varphi_{n}(t) : n,t=0,1,\dots\}\) be independent sets of nonnegative integer-valued random variables defined on a common probability space. The controlled branching process with random control functions is defined by the relationships \[ Z_0=N,\qquad Z_{n+1}=\sum_{j=1}^{\varphi_{n}(Z_n)}X_{nj},\quad n=0,1\ldots . \] Under certain assumptions on the characteristics of the process, the most important of which is the property that the sequence \(\{k^{-1}E\varphi_{0}(k)\}_{k\geq 1}\) converges to a finite value \(\tau \) with \(\tau EX_{01}>1\) in such a way that the sequence \(| \tau -k^{-1}E\varphi_{0}(k)| \) is nonincreasing, the authors establish existence of a random variable \(W\) such that \(\lim_{n\to\infty}Z_n(\tau EX_{01})^{-n}=W\) exists a.s. and describe conditions when one can extend this convergence to convergence in \(L^1\). Besides, they list conditions to be necessary and/or sufficient for the convergence in \(L^1\) to a finite positive and non-degenerate random variable to take place.




Cited in
(24)








This page was built for publication: On the geometric growth in controlled branching processes with random control function

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4819511)