Optimal production-inventory policies in supply chains with information sharing

In recent years information sharing has received broad attention in supply chain management because it can reduce the effects of demand variability and can help participants in supply chains lower their operational costs. By utilizing shared information, businesses can develop more effective production and inventory management policies, resulting in higher profits and better customer service.; The existing literature on information sharing usually assumes a serial supply chain or, when there are more than one retailer in the system, the upstream supplier receives inventory information from all the downstream members. In this dissertation, we first consider a two-retailer supply chain in which the supplier may choose from the downstream partners to establish information sharing. Through modelling the selective information sharing problem as a Markov Decision Process, we characterize the optimal production and inventory policies as state-dependent base-stock policies. An extensive numerical study is performed to examine how system factors affect the benefit of information sharing and the relative value of information from each retailer.; When the downstream retailers' orders have different economic importance to the supplier, a stock rationing problem occurs because the supplier can reserve some stock dedicated to the more valuable retailers. Most of the current research assumes one-level system with unitary order size. This dissertation generalizes the existing research in the context of a two-level supply chain with information sharing. The threshold-type policies are shown to be optimal for the manufacturer's production and rationing decisions. The benefits of the optimal rationing policies are investigated through numerical examples.; We also study the integrated production and maintenance policy in a two-echelon supply chain with a deteriorating facility. The optimal dynamic policy is shown to be very complicated. Thus, we propose the modified echelon-stock policy as a substitute, and develop exact models for evaluating the performance of this policy. A detailed numerical study also demonstrates insights on the benefits of information sharing when maintenance is preemptable and non-preemptable.