An approach for increasing the level of accuracy in Supply Chain simulation by using patterns on input data

Abstract

Setting up simulation scenarios in the field of Supply Chains (SCs) is a big challenge because complex input data must be specified and careful input data management as well as precise model design are necessary. SC simulation needs a large amount of input data -- especially in times of big data, in which the data is often approximated by statistical distributions from real world observations. This paper deals with the question how the model itself and its input can be effectively complemented. This takes into account the commonly known fact, that the accuracy of a model output depends on the model input. Therefore an approach for using techniques of Knowledge Discovery in Databases is introduced to derive logical relations from the data. We discuss how Knowledge Discovery would be applied, as a preprocessing step for simulation scenario setups, in order to provide benefits for the level of accuracy in simulation models.