Mathematical description of a discrete event simulation (DEVS) using factor analytic method

Abstract

Appropriately capturing the logic or results of a detailed simulation through some kind of aggregation to pass up a model hierarchy is a difficult task. This aggregation task involves two main questions: ‘what’ to aggregate and ‘how’ to perform this aggregation. This paper provides a well-defined approach for the ‘what’ portion of the aggregation process for hierarchical simulation models. We start by characterizing a simulation model in a mathematical format. We then use this mathematical description to identify what portion of a lower-level model can be aggregated, or to capture the logical flow of the full lower-level model. A useful mathematical representation of the simulation structure and logic is through the development of a network representation of the model. This, in turn, can be systematically decomposed into smaller sub-networks by performing model decomposition by means of factor analytic methods, which is the main topic of discussion.