Preventive transshipment decisions in a multi-location inventory system with dynamic approach

This paper addresses the preventive transshipment problem in a multi-location inventory system. Decisions are made before demands are observed to prevent future stock out. Both lateral transshipments (LT) and emergency orders (EO) are available in this system. The objective is to figure out the timing and quantity of preventive transshipment decisions. A Markov decision process is formulated to describe the problem. First a decomposition method is used to divide the multi-location problem into different pairs of two-location problems. Then a sorting heuristics combined with backward dynamic programming (BDP) approach and approximate dynamic programming (ADP) approach is developed to solve this problem. Extensive numerical examples are used to demonstrate the effectiveness of this method. Results show that demand distribution at the retailer affects the way transshipment performs. In addition, product's perishability has obvious effects on transshipment decisions.