Combining knowledge-based systems and simulation to solve rescheduling problems

Knowledge-based systems can help to enlarge the application range for simulation. Acting as intelligent front-ends they communicate with the user and provide statistical knowledge. As an example, the paper presents SIMULEX, a prototype decision support system for short-term rescheduling in manufacturing. SIMULEX couples expert systems and simulation to assist the production manager in handling production disturbances. Its core is a job shop simulator that models the plant and evaluates the results of various rescheduling measures. One intelligent front-end (the access system) gathers information about the disturbance and configures the simulation experiments. Another (the run time control) supervises the simulation runs. A third one (the termination/interpretation system) analyzes the data, derives conclusions and suggests the most promising measures. The paper focuses on the system's decision model that allows multi-attribute decisions under uncertainty.