Quantitative computer simulation as a paradigm of scientific investigations

Abstract

Advances of computer technology initiated in the twentieth century have resulted in adoption of computer simulation as the most popular tool of performance evaluation studies of such complex stochastic dynamic systems as e.g. modern multimedia telecommunication networks. Such wide-spread reliance on simulation studies raises the question of credibility of their results. This question needs to be answered before computer simulation can be objectively accepted as an independent, self-sufficient paradigm of scientific investigations. In this talk, having briefly overviewed the main necessary conditions of any trustworthy simulation study conducted for performance evaluation of stochastic dynamic systems, we will focus on simulation studies with on-line output data analysis. The perils and pitfalls of quantitative discrete-event computer simulation will be considered, together with its fast distributed version, known as Multiple Replications in Parallel.