Evaluation of Indicators for Simulation's Prediction Quality of Material Demand in Matrix Production Systems

The resilient and relatively new production system "matrix production"challenges the design of the material supply. Through its dynamic behavior, the time, place and amount of material demand is uncertain. To support the supply planning process, it takes a method that predicts the material required in a given planning period. The usage of simulation can clarify the expected material demand in a given period. To evaluate the simulation's prediction quality and set a limit to the simulation experiments, an indicators or goodness-of-fit value is needed. This article examines different statistical indicators with regard to their suitability for evaluating the prediction quality of a simulation and its multiple experiments. For this purpose, a use case is investigated to analyse how the indicators change over a number of experiments. The indicators are evaluated and discussed with respect to their suitability in further usage.