| A goal directed simulation environment for discrete part manufacturing systems is developed. The first of the research involved the identification of the expert knowledge that an analyst may use during analysis of different types of manufacturing systems. Based on this expert knowledge, a four level structure for the rule base was developed. Level 0 is comprised of rules that make the final decisions as to what changes will be made to the model parameters. Level 1 rules are mainly responsible for identifying what changes could be made depending on the goals and the current status. Level 2 rules determine what Level 1 rules may be used, depending on the type of manufacturing system being analyzed. In addition to these, a fourth level is comprised of information specific to the system being analyzed. A general set of rules was then developed to demonstrate that this structure can handle a wide variety of situations.;The second phase of the research was to develop a more detailed set of rules. The study focussed on a limited set of manufacturing systems, and decisions were limited to only capacity changes, with the purpose of satisfying three types of goals: (1) increasing throughput to obtain a minimum desired throughput, (2) reducing tardiness to maintain the percentage of jobs that are tardy, below a desired level, and (3) reducing work-in-process so that it is below a desired level. The rules were developed iteratively using different models of manufacturing systems. On comparing with a modified version of Hooke-Jeeves method, the rule based approach was able to give a solution in lesser number of iterations, in all the examples considered, demonstrating the utility of such an approach.;The third phase was the implementation of a prototype goal directed simulation environment (GDSE) in a microcomputer environment. GDSE is a hybrid system built around the SIMAN simulation language, and makes use of the rules developed in the second phase. The expert system shell used for the purpose was CLIPS (C Language Integrated Production System). |
