Critical branching processes in a sparse random environment (Q6067092): Difference between revisions

The authors study critical Bienaymé-Galton-Watson processes \((Z_n)_{n\ge 0}\) assuming a random environment not for all generations but only for those marked by positions of a random walk along the sequence of generations: Let \((d_n)_{n\ge 1}\) be a sequence of positive integer valued i.i.d. random variables and \((\nu_n)_{n\ge 1}\) a sequence of i.i.d. random probability measures on \(\mathbb{N}_0\), independent of \((d_n)_{n\ge 1}\). Define the random walk \(S_0:= 0\), \(S_k:= \sum^k_{j=1} d_j\), \(k\in\mathbb{N}\). Then \(Z_0= 1\), \(Z_{n+1}= \sum^{Z_n}_{j=1} \xi^{(n)}_j\), \(n\in\mathbb{N}_0\), where the \(\xi^{(n)}_j\) are independent random variables, independent of \(Z_n\), and with common distribution \(\nu_{n+1}\) whenever \(n=S_k\) for some \(k\), and with some common non-random distribution for all other \(n\). Let \(\mathbf{m}:= \mathbf{E}d_1<\infty\), \(\mathbf{v}^2:=\text{Var}(\log \sum^\infty_{j=1}j\nu_1(j))\in(0,\infty)\), and suppose \((Z_n)_{n\ge 0}\) is critical. Then, imposing some additional moment conditions, the authors show that, as \(n\to\infty\), \(n^{1/2}\mathbf{P}(\text{inf}\{k\ge 0: Z_k= 0\}>n)\) tends to a finite positive number, and the law of \((\log Z_{[nt]}/\mathbf{v}(n/\mathbf{m})^{1/2}|Z_n>0)|_{t\in[0,1]}\) tends weakly in the space of positive measures on \(D[0,1]\) with Skorokhod \(J_1\)-topology to the law of the Brownian meander \(B_+(t):= (1-\zeta)^{-1/2}|B(\zeta+t(1-\zeta))|_{t\in[0,1]}\), where \(B(t)\) is standard Brownian motion and \(\zeta:= \sup\{t\in[0,1]: B(t)= 0\}\).