A MODEL FORMALISM FOR THE DESIGN OF SIMULATION MODEL REPRESENTATION LANGUAGES (ABSTRACT DATA TYPES, MACHINE, SPECIFICATION)

PETER HUICHUN CHEN, Purdue University

Abstract

The methodology of simulation has advanced substantially over the past two decades. However, it has not been widely accepted by the production decision-making community yet. One of the major reasons identified for this situation is the inability to derive supportive simulation analysis results within the decision-making time frame. And, this is, in turn, caused by ineffectiveness and inefficiency in simulation model development. To solve this problem, recent research efforts have suggested the development of a Simulation Model Management System (SMMS) to provide an integrative environment to support model construction/implementation, model experimentation, result analysis, and model maintenance. To successfully develop such a system, however, requires a knowledge of the properties of simulation models. In this research, we limit our considerations to the models employed in discrete event simulation analyses. A model formalism is developed to provide a formal basis for investigating the properties of such models. The emphasis of this development is primarily concentrated on the formal support in the design of an appropriate tool--model representation language, for the computer implementation of discrete event simulation models. In developing this model formalism, the production system concept proposed by Post is adopted as a framework. Some of the basic discrete event system concepts discussed in Zeigler's formalism work are also incorporated. Meanwhile, in the formalism a set of abstract data types are derived to further formalize the operational characteristics presented in the structure and behavior of discrete event simulation models. The main contribution of this research is a formal framework for analyzing language features' expressiveness during the design of a model representation language. Other contributions are in providing: (1) a set of notations for deriving a formal model description of a simulation problem and (2) a formal way of specifying the semantics of discrete event network simulation languages. In order to demonstrate the value of these results, examples are given at the end of the thesis.

Degree

Ph.D.

Subject Area

Industrial engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS