On limit distributions of the time of first passage over a high level

Abstract

We describe a simple method of finding the limit distribution of the time of first passage over a high level for a random sequence (or random process). The method reduces the considered problem to a well-studied problem of finding the distribution of this random sequence (or process) at a certain fixed time. This allows to consider not only traditional sums of independent random variables but also sums of dependent random variables. As examples, sums of independent random variables with stable limit distribution functions, sequential testing of hypotheses, and problems of the first passage over a high level for Gaussian processes with strong dependencies are considered.