A coordinated manufacturer-retailer model under stochastic demand and production rate

This paper deals with a manufacturer-retailer model for a two-stage supply chain. The retailer and the manufacturer face random demand and random yield, respectively. Each time the manufacturer switches from an idle mode to a production mode a fixed cost is incurred. Inventories can be deployed both at the manufacturer and the retailer. This description of a two stage system is more detailed than generally seen in the literature. We first develop a Markov chain model to compute the optimal coordinated decision policy. This model serves as benchmark to study the performances of three non-coordinated models. In all three models the retailer first computes an optimal (s,S) policy which the manufacturer then comply with. The difference among three models is how the backlogs at the manufacturer are handled. The numerical results show that the non-coordinated model in which the backlog is immediately canceled generally performs best.