Weak convergence of random sums and maximum random sums under nonrandom norming (Q1354965)

Necessary and sufficient conditions are obtained for the weak convergence of random sums and maximum random sums of i.i.d. random variables. Limit distributions for these sums are described. Let \(v\), \(N_n\), \(X_n\), \(n=1,2,\dots\), be random variables defined on a probability space \((\Omega,{\mathcal A},P)\), where \(v\) is positive with probability one, each \(N_n\) takes natural values, and \(X_n\), \(n=1,2,\dots\), are i.i.d. with \(EX_k=a\neq 0\) and \(\text{Var }X_k= \sigma^2\), \(0<\sigma^2<\infty\). Denote \(S_k= X_1+\cdots+X_k\) and let \[ Z_n={S_{N_n}-avn\over\sqrt{\sigma^2n}},\quad \overline Z_n={\max_{1\leq k\leq N_n} S_k-avn\over\sqrt{\sigma^2n}}. \] The random variables \(Z_n\) and \(\overline Z_n\) are called a random sum and a maximum random sum of random variables under nonrandom norming, respectively. The results are the following: Theorem 1. In order that \(P(Z_n<x)@>d>>F(x)\) for some distribution function \(F\) it is necessary and sufficient that \[ P\Biggl({S_{[vn]}- avn\over\sqrt{\sigma^2n}}+ {a(N_n- vn)\over\sqrt{\sigma^2n}}< x\Biggr)@>d>> F(x).\tag{1} \] Theorem 2. Assume that \(EX_k=a>0\). In order that \(P(\overline Z_n<x)@>d>>F(x)\) for some distribution function \(F\) it is necessary and sufficient that (1) holds.