Proceedings of the Seventh Int’l Conf. on Artificial Intelligence and Expert Systems Applications, San Francisco, November, 1995. ENHANCING OBJECT-ORIENTED SIMULATION WITH RULE-BASED EXPERT SYSTEM Otto Lee & Hon Wai Chun City University of Hong Kong Department of Electronic Engineering Tat Chee Avenue Kowloon, Hong Kong email: eehwchun@cityu.edu.hk ABSTRACT This paper describes our research in developing an intelligent simulation system by enhancing an object-oriented simulator with a rule-based expert system. Designing and implementing a simulation model for a particular problem requires extensive expertise in simulation techniques, computer languages and queuing theory. However, those who would like to use simulation to assist their decision making might not always be technically knowledgeable enough to develop simulation programs. The primary motivation for our research is to shift the burden of simulation software development to an intelligent artificial intelligence system. The test domain for our research is simulating the queuing situation at airport check-in counters. The main objective for the simulation is to determine how many counters are needed in order to check-in all passengers given a certain level of service standard. KEYWORDS knowledge-based simulation, resource allocation, scheduling, expert system 1. INTRODUCTION This paper describes an intelligent domain-specific simulator called RBOOS. This intelligent simulator was developed as an experiment in integrating an object-oriented simulator with a standard OPS5 rule-based expert system to provide a more accurate and user friendly approach to performing simulation experiments. RBOOS is an airport simulation system which can predict the number of check-in counters needed to adequately service the passengers of one departing flight. The graphic user interface was designed so that a novice user can easily use the simulator without technical training in simulation techniques. RBOOS will analyse the current problem (defined as a set of flight information), determine appropriate simulation parameter values using the rule-based expert system, build the simulation model, run the simulation, compare the simulation output with the desired goals, make changes to the model parameters if necessary, and continue until the simulation goals are achieved. The simulation goals define the level of service required at this airport and is measured in terms of queue length and waiting time. RBOOS consists of four main components (see Fig. 1) - the Object-Oriented Simulation Module, the Animated Graphics Module, the Intelligent Front-End Module, and the Intelligent Back-End Module.