COMPUTER SIMULATION OF QUEUES

Abstract

Imulation Is poriormlnk centrolled experiment on h modelL to study the eharaeteristics or the system that modoi represents. It Is •a powerful technique for solving a wide variety of problems. Simulation may be numerical sequential or electrical or physical according to the method of experiment. Obviously, computer simulation is a numerical sequential one. If one could get a firm handle on how queues work, he would be able to manage them better and perhaps even eliminate them. Simple waiting lines can be analyzed mathematically, but most queueing situations are so complex that they defy precise descriptions. For these situations, computer can be used to model and simulate a queueing situation to make predictions about the queue and to learn how it behaves. In the present dissertation, a simulation program "SIMQ" has been developed in Turbo C on IBM PC/XT(AT) compatible, which can simulate many queueing situations. SIMQ uses Monte Carlo simulation technique. The queueing systems, single channel system, multiple channel,single queue system and multiple channel' multiple queue system can be simulated using SIMQ. The system may have finite or infinite queues. Queue dispatching allowed is first come, first served. The program supports exponential, Erlang, normal, uniform probability distributions and also deterministic case for service time and inter-arrival time. Further, it allows the user to have his own probability distribution other than these standard ones..