Ranking a list of discrete-event models

Abstract

It has been demonstrated that Arena simulation models can be parameterized and stored in a relational database for later use. In this way an Arena model can be uniquely described by assigning the right values to the corresponding parameters. These parameters contain detailed information concerning type and number of modules, entities and routes which entities follow through the system. A modeler could query such a database hoping that it would contain the Arena (sub)model that one has in mind. Unfortunately, an exact match will seldom occur but the database might contain one or more similar models. In earlier work a numerical algorithm was proposed resulting in a ranked list of simulation models in the database. According to this the lower a model is ranked the more it resembles the specified model. This model depends on a number of weight factors. In this paper the ranking algorithm is evaluated by carrying out a number of experiments.