A Bayesian Neural Network Approach to Synthesising Data for Supply Chain Optimisation and Simulation Modelling

Abstract

The Bayesian learning model developed in this research addressed a need for technology by which complex industry datasets may be synthesised and represents a fundamentally novel approach to this problem than those previously explored in literature. This approach enables the characterisation of complex stochastic phenomena from historical and current observations under the influence of industry forecasts, to yield datasets that are sufficiently lifelike for use in optimisation and simulation modelling, when real data is scarce.