Discrete event simulation is an important system analysis technique. But for today’s demand for speed, the time to complete a simulation study is considered to be long, even with current developments in hardware and simulation software. In this scenario, simplification methods for simulation models could play a key role. This paper proposes a technique for reducing the complexity of a discrete event simulation model at the conceptual phase of simulation modeling that can be fully automatized through a computer program. We applied this technique on some problems which demonstrate the feasibility of this approach.

Simulation model representation technique

In order to apply a model reduction procedure the Simulation Model should be represented in a proper Simulation Model Representation Technique. There are several techniques for simulation model representation found in the literature regarding discrete-event systems. We could divide these techniques into two categories:

diagrammatic (or graphical) methods and formal specification languages.

As graphical languages, we can cite Activity Cycles Diagrams [30–32], Control Flow Graphs [33], Event Graphs, Petri Nets [35] and its extensions [36], State Charts [37] and Process Networks [38]. As formal specification languages, one of the most cited is Discrete Event Systems Specification (DEVS) from [20], based on the systems theoretic approach. Another example is the Condition Specification (CS) created by Overstreet and Nance [39], based on the ‘‘Conical Methodology’’.By analysing several model representation techniques we found that the one suitable to a simplification technique due to the ability to represent all model details is the Condition Specification. However this representation is difficult to construct from scratch ([40] also pointed out that the Condition Specification demands some buffering mechanism between the representation and the analyst). So we will consider the Activity Cycle Diagram a ‘‘higher level modelling’’ to be build first and then after to be converted into Condition Specification.

To allow this conversion, we have developed a series of automatic rules to convert ACD into Condition 932 L. Chwif et al. / Simulation Modelling Practice and Theory 14 (2006) 930–944 Specification. It is not scope of this paper to cover these conversion rules. More detail is provided in [41]. Appendix A reviews the basic principles of the Activity Cycle Diagrams and the Condition Specification. In the next section, we will be presenting our reduction algorithm based on the defined reduction problem to be applied on models represented into Condition Specification format.