scientific article; zbMATH DE number 4013643

From MaRDI portal
Publication:3025960





Let \(\xi_{in}\), \(k=1,2,...\), be independent identically distributed random variables and \(\nu_ n\) a random variable with values 0,1,2,... and independent of \(\xi_{in}\), \(i=1,2,... \). Define the sum \[ T_ n^{(r)}(\nu_ n)=\sum_{1\leq i_ 1<...<i_ r\leq \nu_ n}g_ n(\xi_{i_ 1n},...,\xi_{i_ rn}) \] where \(g_ n(\cdot)\) are symmetric functions with values 0 and 1 such that: 1) the sequence of random variables \(\{C^ r_{\nu_ n}E g_ n(\xi_{1n},...,\xi_{rn})\}\) converges weakly to a random variable \(\nu\), as \(n\to \infty,\) 2) \(\ell_ n^{-2r}E[g_ n(\xi_{1n},...,\xi_{rn})g_ n(\xi_{1n},...,\xi_{(r-1)n},\xi_{(r+1)n})]\to 0\) as \(n\to \infty\) if \(\lim_{n\to \infty}C^ r_{\ell_ n}E g_ n(\xi_{1n},...,\xi_{rn})=\lambda >0.\) The main result is that: \[ \lim_{n\to \infty}P(T_ n^{(r)}(\nu_ n)=k)=(1/k!)\int^{\infty}_{0}x^ ke^{-x}dP(\nu \leq x) \] if conditions 1) and 2) hold. This assertion is used in some problems for branching processes.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3025960)