Inventory optimization in a retail multi-echelon environment

Abstract

The objective of the study is to find an optimal inventory distribution in a retail three-echelon environment, consisting of a supplier, a DC, and stores. An inventory model is built by replicating the echelons' periodic, order-up-to-level policies with all echelons' transactions integrated. Network carrying cost is set as an objective function, while the store target service level and the store's minimum order-up-to-levels are set as constraints. A heuristic approach, that combines the optimization and simulation methods, is used to find the optimal inventory distribution. The results show that the optimal network carrying cost can be achieved by having low inventory and low service level at the DC. In addition, the impact of the echelons' deviations from the optimal policies as well as the impact of the upstream echelon's service disruptions on the other echelons confirms the interrelation between the echelons in the network. The analyses also illustrate that high target service level can be accomplished by keeping high inventory at the stores and the DC.