On the relative merits of two major methodologies for simulation model construction

• Authors: Charles M. Shub
• Venue: WSC '78 Proceedings of the 10th conference on Winter simulation - Volume 1
• Year: 1978

Given a model of a discrete change system there are two major methodologies for developing a simulation program of the model. They are (A) describing the flow of typical units of traffic through the model, and (B) describing what occurs at those instants of time when the state of the system changes. This paper compares, without entering into the controversy of which language is best for modeling, the relative advantages and disadvantages of the two methodologies. The comparison is made at the modeling or source language level of detail and does not specifically adjudicate the relative merits of language package internals such as dynamic storage implementation, the notion of an events chain, be it current or future, or the notion of a timing routine except as necessary for comparisons. The paper is concerned with the relative advantages of the two methodologies in handling such problems as simultaneous occurrences, synchronization of traffic units, communication between traffic units, complex decision algorithms involving waiting line manipulation, the general notion of “style” in the code produced, the ease of gathering meaningful statistics, the level of knowledge required of the system internal structure, and the refinement of level of detail.