Synthetic Demand Generation with Seasonality for Data Mining on a Data-Farmed Data Basis of a Two-Echelon Supply Chain

Abstract

A widely used method in the context of supply chain analytics and management is data mining. It is used to discover patterns in a supply chain's data basis. Besides preprocessing observational real-world data, simulation can be used to create a suitable data basis. This process is referred to as data farming and involves using a simulation model as a data generator. A common input to a simulation model of a supply chain is demand of stock keeping units that is likely to project to the model's behavior. When testing novel approaches or in planning stages, demand of real-world supply chains is not always available or viable to adept. In this case, synthetically created demand can be used. A general approach of realistic demand generation with seasonality by a demand generator in the context of data farming is presented and further exemplified on a data farming and data mining framework of a two-echelon supply chain.