Limit theorems with refinements for random processes (Q1120896)

[For the entire collection of the original see Zbl 0626.00025.] The goal of the paper is to indicate a variety of the results on limit theorems with refinements in functional spaces. Attention is paid to representations of the following type: \[ M f(\xi^{\alpha})=M[f(\xi_*)+\epsilon (\alpha)\sum^{k_ 1}_{j=1}\int_{[t_ 0,T]^ j}\quad f^{(j)}(\xi_*;ds_ 1,...,ds_ j)\Gamma^{1,j}(s_ 1,...,s_ j)+...]+o(\epsilon (\alpha)^ s\quad), \] where \(\xi^{\alpha}\) is a family of random processes on the segment \([t_ 0,T]\) and \(\xi_*\) is the limit process for \(\xi^{\alpha}\), \(\epsilon(\alpha)\to 0\) (when \(\alpha\) tends to some limit), f belongs to some smooth class of functionals, and \(\Gamma^{1,j}(\cdot)\) are certain functions. One can use these results to compute M \(f(\xi)\) for random processes \(\xi\) by the Monte Carlo method. The analogies with the classical scheme of investigation of sums of independent random variables are indicated, and results for the classes of Markov and Gaussian processes are given.