Supply chain hybrid simulation: From Big Data to distributions and approaches comparison

Abstract

The uncertainty and variability of Supply Chains paves the way for simulation to be employed to mitigate such risks. Due to the amounts of data generated by the systems used to manage relevant Supply Chain processes, it is widely recognized that Big Data technologies may bring benefits to Supply Chain simulation models. Nevertheless, a simulation model should also consider statistical distributions, which allow it to be used for purposes such as testing risk scenarios or for prediction. However, when Supply Chains are complex and of huge-scale, performing distribution fitting may not be feasible, which often results in users focusing on subsets of problems or selecting samples of elements, such as suppliers or materials. This paper proposed a hybrid simulation model that runs using data stored in a Big Data Warehouse, statistical distributions or a combination of both approaches. The results show that the former approach brings benefits to the simulations and is essential when setting the model to run based on statistical distributions. Furthermore, this paper also compared these approaches, emphasizing the pros and cons of each, as well as their differences in computational requirements, hence establishing a milestone for future researches in this domain.