A method for considering aleatory and epistemic uncertainties as well as product variance in discrete event simulation of production systems

When modelling a production system during its planning phase, aleatory uncertainties of production processes, epistemic uncertainties resulting from insufficient knowledge as well as variations in the production processes resulting from product variances must be considered. These different uncertainties and variances are interrelated, e.g. the influence of product variants on production processes may itself be subject to epistemic uncertainty. This paper presents a generic method to model aleatory and epistemic uncertainties in discrete event simulations of production systems as well as product variances in an integrated manner. We use functional relations between product parameters and production model parameters to efficiently account for product variances. We use possibility-probability transformation and second-order Monte Carlo simulation to account for epistemic uncertainty. For easy transferability to industrial practice, a step-by-step procedure is described that can be implemented in commercially available simulation tools. A use case from precast concrete production is presented to show the benefit of such an approach compared to a state-of-the-art benchmark.