Two essays on optimal procurement policies in supply chains with constraints.

There are two sources of complexity when dealing with production or distribution systems: an exogenous source, related to the variations in the environment surrounding the firm; and an intrinsic source, related to the inherent characteristics of the product or process. Two inventory control problems that include an exogenous factor and an intrinsic factor, respectively, are presented. In the first problem, we consider a periodic-review system with two priority demand classes: one deterministic, from contracted customers; and the other stochastic, from occasional customers. Units of stochastic demand not satisfied when demand occurs are treated as lost sales. One has to decide whether an order should be placed, how much to order, and how much demand to fill from the stochastic source. We characterize the structure of the optimal policy and show that the optimal order and rationing quantities are state dependent. A simple (s, k; S) policy is proposed, where s and S determine when and how much to order while k specifies how much of the stochastic demand to satisfy. In the second problem, we consider a periodic-review system of an item with two components, Part 1 and Part 2. While the life time of Part 1 is long, Part 2 is perishable and only lasts one period. One has to decide whether an order should be placed and how much of each item to order. The items are coupled and the ordering decisions have to be determined simultaneously. Considering a fixed ordering cost for Part I and a backorder cost for demand not satisfied from stock, we characterize the structure of the optimal policy as an ( sn, S1n , S2n ) policy, where sn and S1n specify when and how much of Part 1 to order while S2n determines how much of Part 2 to order in period n. This dissertation makes a substantial addition to the area of inventory control by providing optimal control policies for complex problems. Failing to consider these new characteristics and applying traditional policies can significantly affect the firm's profitability. Moreover, the insight gained while characterizing these policies will be useful to study systems with an even higher level of complexity.PhDApplied SciencesIndustrial engineeringOperations researchUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/123171/2/3068864.pd