Analysis of three optimization problems in operations management

This dissertation addresses three novel problems in both manufacturing and service operations management.;The first problem focuses on the service operations problem related to the subscriber-based online movie rental through a DVD-by-mail service. We must understand subscribers' demands due to rapidly-declining demands. We accordingly develop a demand formulation based on data from Blockbuster. Using the demand formulation derived, we construct an optimization problem that makes jointly two decisions: (i) the initial order quantity of a new movie title, and (ii) its shipment sizes of the title for every period after its release. We then develop a simple policy that is to ship as large an amount of DVDs as possible to subscribers. This policy is shown to be potentially optimal through extensive computational experiments.;In the second problem, we study the challenging problem that improves the efficiency of supply chains with reverse logistics. We consider a multi-period two-product supply planning problem in a manufacturing system with two stages. The remanufacturing stage is the first stage that provides two closed-related components to the second (manufacturing) stage, which uses each component to manufacture new products. Our objective is to develop optimal production plans that minimize the total cost at the remanufacturer for various production strategies. We then establish the complexity status for each strategy, and comprehensive computational experiments provide insights into this closed-loop supply chain for those strategies.;The third problem investigates the throughput optimization in dual-gripper bufferless robotic cells with a circular layout. The cell is designed to produce identical parts under the free-pickup criterion with additive travel time. We attempt to obtain an efficient algorithm for an optimal k -unit cyclic solution (k ≥ 1) with the objective of maximizing the throughput. We devise polynomial algorithms that offer an approximation solution of the worst-case bound for single-gripper and dual-gripper robot cells. Through computational experiments, we demonstrate that the algorithm performs much better on average than this worst-case bound suggests. We believe that our results will provide managerial insights for scheduling operations that maximize productivity for any configuration of the cell data or for either type of robot.