Expert systems and simulation

Up to now, it has not been possible to incorporate the expert knowledge necessary for the overall modelling within an expert system which would then in turn make the techniques accessible to the inexperienced user. In general, the close cooperation among experts in the fields of application and simulation techniques remains-for the time being-the only guarantee for successfully scrutinizing a model study. It is, however, possible to combine expert systems and simulation models when using carefully chosen, closely defined fields of application. It is here that an important area of interest and study lies. There are three conceivable alternatives: (1) To provide a simulation model with an intelligent environment which supports the user in constructing as well as evaluating the results made. (2) An expert system contains a simulation model as a possible component. If the expert system needs information that is not directly available, a simulation model is accessed which in turn sends the necessary data back to the expert system. (3) A simulation model contains an expert system as a component. In this case, the model has specific strategies at its disposal that are to be taken into consideration. If the model reaches a point where decisions have to be made, the simulation model turns to the expert system in order to receive information on relevant decisions. These relevant decisions are then used to determine the further course of the model. All three alternatives assume that the expert system as well as the simulation system have an open architecture. The open architecture is a prerequisite for the possibility of coupling the two. Up to now, this requirement has rarely been achieved.