The joint transshipment and production control policies for multi-location production/inventory systems

Abstract

In this paper, we study the joint transshipment and production control policies for multi-location production/inventory systems in which items are manufactured and stocked at each location to meet incoming demand. We formulate the problem as a make-to-stock queue to gain insight into the following questions: (1) How much demand at a location should be covered by transshipment from other locations, and when to produce or stop production? (2) Is there a simple structure associated with the optimal policy, and whether a simple decision rule can be implemented for transshipment control? (3) Can effective heuristic policies be developed to solve the multi-location problems? For the two-location problem, we characterize the optimal policy as monotone switching-curve policy. To address the multi-location problem, we develop two heuristic policies. One is obtained from the one-step policy improvement based on policy iteration and the other from the one-step lookahead method based on the approximation of the optimal cost function. Numerical examples are used to illustrate the optimal and heuristic policies and compare their performance for various cases.