Abstract

Simulation is a much maligned and misunderstood methodology. To be sure, not all applications of simulation are successful. But is this the fault of the methodology or those who attempt to make use of it? Much of the criticism of simulation is a result of the fact that critics expect too much of a simulation model they expect the model to be both a tool for prediction and a tool for the analysis of management systems. These are two separate problems entirely and they must be treated as such in the development, evaluation, marketing and use of a simulation model. Economists provide a representative example of conflicts that can arise when using simulation. They often develop models for the purpose of predicting economic variables several quarters into the future. Each relationship in the model is based upon empirical evidence that is gathered and analyzed by appropriate statistical techniques. Statistical significance is the ultimate criterion in developing each relationship. These models are tested against known data and adjustment factors are introduced to "tune" the model. Various models are compared with respect to the accuracy of their predictions from quarter to quarter.