Maxima of random particles scores in Markov branching processes with continuous time (Q2271710): Difference between revisions

A Markov branching process \(Z\) with continuous time is considered, in which each particle has a univariate or bivariate score. The scores of different particles are i.i.d. with CDF \(F\). The behaviour of the maximum \(M(t)\) of the scores is investigated as the time \(t\to\infty\). E.g., in the bivariate case \(M(t)=(M_1(t),M_2(t))\) we have \[ P(M_1(t)<\alpha_1(e^{\kappa t})x_1+\beta_1(e^{\kappa t}), M_2(t)<\alpha_2(e^{\kappa t})x_1+\beta_2(e^{\kappa t}))\to \varphi(-\ln G(x_1,x_2)), \text{ as } t\to\infty, \] if for some \(\alpha_i(r)\), \(\beta_i(r)\), \(G(x_1,x_2)\) \[ F^r(\alpha_1(r)x_1+\beta_1(r),\alpha_2(r)x_1+\beta_2(r))\to G(x_1,x_2), \text{ as } r\to\infty, \] where \(\varphi(s)=E e^{-sW}\), \(W=\lim_{t\to\infty}Z(t)e^{\kappa t}\), \(\kappa=\lambda(\mu-1)\), \(\lambda\) is the intensity of births, \(\mu\) is the mean of offsprings number for \(Z\). The class of possible limit distributions is investigated.