Evaluation of Forecasting and Inventory Control in Multi-product Manufacturing Systems Operating under Erratic Demand: A Case Study in the Automotive Domain

Erratic demands have traditionally shown complexity in inventory management and forecasting. Several organisations, especially in multi-product manufacturing systems with intermittent demand high level of variability and uncertainty, assume suitable estimators are put in place to forecast demand. This forces them to pay attention to controlling the inventory of a system. The forecasted demand is often characterised by one or more (hybrid) statistical distributions. The latter, often used in modelling highly variable intermittent demand profile, has a significant effect on the inventory and service level in a system. This paper evaluates the demand distributions frequently used for inventory control under erratic demand variability. We use a discrete event simulation technique to analyse the level of significance of statistical distributions. We then prove our methodology on real industrial data of a multi-product manufacturing plant. We aim to determine the most suitable approach for inventory control with respect to minimising inventory while maintaining zero stock-outs and high service level.