Single-Stage Approximations of Multiechelon Inventory Models

Abstract

This chapter summarizes recent development on simple heuristics for optimal inventory policies of multi-echelon inventory systems. These simple heuristics are based on solving a sequence of single-stage inventory problems whose parameters are obtained from the original system data. We mainly focus on series systems under continuous-review and periodic-review schemes and briefly discuss the key results for the assembly and distribution systems.

The information is centralized, and the objective is to minimize the average total cost per time period/the expected discounted cost over a planning horizon. When fixed order costs are negligible, echelon base-stock policies are considered. On the other hand, when fixed order costs are significant, we consider (r, q) policies for the continuous-review system and (s,T) policies for the periodic-review system. These heuristics not only simplify the computation and implementation, but also help gain insights into managing inventory in supply chains. We also provide a summary of how to obtain simple solutions to some extend models.

Keywords: Multiechelon inventory systems, single-stage approximations, echelon base-stock policy, newsvendor solution, (r,q) policy, (s,T) policy