Large deviations for weighted random sums

Abstract

In the present paper we consider weighted random sums Z_N = ∑_j=1^Na_jX_j, where 0 ≤ a_j < ∞, N denotes a non-negative integer-valued random variable, and {X, X_j , j = 1, 2,...} is a family of independent identically distributed random variables with mean EX = µ and variance DX = σ² > 0. Throughout this paper N is independent of {X, X_j , j = 1, 2,...} and, for definiteness, it is assumed Z₀ = 0. The main idea of the paper is to present results on theorems of large deviations both in the Cramér and power Linnik zones for a sum ^~Z_N = (Z_N − EZ_N )(DZ_N )^−1/2 , exponential inequalities for a tail probability P(^~Z_N > x) in two cases: µ = 0 and µ ≠ 0 pointing out the difference between them. Only normal approximation is considered. It should be noted that large deviations when µ ≠ 0 have been already considered in our papers [1,2].