AN ENHANCER FOR DISCRETE EVENT SIMULATION LANGUAGES

SIMULATION LANGUAGES, SUCH AS GPSS/H, SIMSCRIPT II.5 GASP, ETC. ARE BECOMING INCREASINGLY POPULAR AS THE TOOL IN THE DETAIL DESIGN AND ANALYSIS OF DISCRETE EVENT SYSTEMS, SUCH AS TRANSPORTATION NETWORKS, MANUFACTURING SYSTEMS, COMPUTER/COMMUNICATION NETWORKS. THE BASIC LIMITATION OF SIMULATION IN REAL WORLD APPLICATIONS STUDIES IS TIME AND COST (TIME TO CONSTRUCT AND RUN THE PROGRAM, COST OF EXTENSIVE ITERATIONS). PARAMETRIC STUDY OF A SYSTEM VIA SIMULATION IS THUS EXPENSIVE AND TIME CONSUMING. WE PROPOSEIN THIS STUDY THE RESEARCH AND DEVELOPMENT OF A GENERAL PURPOSE SIMULATION ENHANCE TO AN EXISTING SIMULATION LANGUAGE USING THE RECENTLY DEVELOPED TECHNIQUE OF PERTURBATION ANALYSIS (P/A). THE IDEA IS THAT WITH THE ENHANCER THE P/A THEORY WILL BECOME TRANSPARENT TO THE PROGRAMMER. WITHOUT ANY THEORECTICAL P/A KNOWLEDGE ON HIS PART, THE SIMULATION PROGRAM HE WRITES WILL AUTOMATICALLY PRODUCE SENSITIVITY OR "WHAT IF" ANALYSIS INFORMATION SIMULTANEOUSLY WITH REGULAR SIMULATION OUTPUTS IN A SINGLE EXPERIMENT. TYPICALLY THIS WILL RESULTS IN A N-FOLD INCREASE IN PRODUCTIVITY OF THE SU.IMULATION WHERE N IS THE NUMBER OF "WHAT IF" QUESTIONS ASKED. MORE IMPORTANTLY, BECAUSE OF THE IMMEDIATE AVAILABILITY OF THE SENSITIVITY INFORMATION WITH THE REGULAR SIMULATION OUTPOT, TURN AROUND TIME IN DESIGN AND ANALYSIS WILL SIGNIFICANTLY DECREASE AND THUS INCREASING THE PRODUCTIVITY OF THE ANALYST.