The Impact of Correlated Demands on Integrated Safety Stock Optimization in Multi-Stage Manufacturing Systems

Abstract

In this paper the determination of safety stocks in multi-stage manufacturing systems with — at least partly — serial and divergent structures is considered in case of stochastic demands which are correlated both between products and also through time. It is shown that these types of correlation have contrary effects on the safety stock distribution, and that neglecting demand correlation can lead to significant deviations from the optimal buffer policy. A procedure for integrated multi-level safety stock optimization is presented which can be applied to arbitrary serial and divergent systems, even if demand is jointly correlated.