A stable iterative algorithm for evaluation of upper and lower approximations to ultimate ruin probabilities

Abstract

In this contribution we present an iterative algorithm for the calculation of two-sided numerical approximations to the probability of ultimate ruin for the classical risk model, which can be directly used in the case of compounding assets as well. Examples involving claim size distributions for which the existing recursive algorithms are inherently unable to provide numerical upper and lower bounds are illustrated, showing the merits of the proposed approach.