Comparison of supply chain planning concepts for general multi-item, multi-echelon systems

In this paper we present two alternative concepts to control general multi-item, multi-echelon systems under stochastic stationary demand for end items. Such systems consist of items that are assembled from other items and in turn are assembled into other items. Each assembly process involves a planned lead time. For such systems optimal control policies are unknown to-date. Therefore we resort to control concepts that at least enable an exact computation of the control parameters. The two alternative concepts represent two fundamentally different modeling concepts: an LP-based concept representing application of deterministic mathematical programming models in a rolling schedule context (the common practice in so-called Advanced Planning Systems), and modified base stock policies representing application of classical multi-echelon inventory models. The parameters of the LP-based concept can be detennined by discrete event simulation. The parameters of the modified base stock policies can be determined analytically. We compare the two concepts based on the required supply chain capital investment required to guarantee target end item service levels. Surprisingly, the modified base stock policies outperform the LP-based concept. We provide managerial insights as well as a deeper understanding into a number of fundamental issues related to supply chain planning and supply chain design.