Essays on optimization and system design in service and supply chain management

This dissertation is comprised of three essays that model, analyze and optimize service and supply chain systems in the context of some specific issues. In the first essay, we help optimize a service system by addressing the question of whether prioritization is helpful. There are two classes of customers, urgent with a high waiting cost and non-urgent with a low waiting cost. There are two types of employees, sorters who collect information on a customer and then decide whether it is urgent or non-urgent, and processors who provide the service. Therefore, prioritization costs both time (additional waiting to get prioritized) and money (salary paid to sorters). We compare sorter-plus-processor and processor-only systems under different scenarios to answer this question.In the second essay, we analyze whether a supply chain system is performing optimally or not. We do it by studying supply chain coordination in a newsvendor scenario when the inventory is managed by the vendor (VMI). We consider four popular contracts from the literature, buyback, quantity flexibility, quantity discount and sales rebate contracts, and analyze their ability to coordinate the supply chain under VMI. Since all of them do not coordinate, we propose two new contracts that coordinate under VMI and analyze their ability to coordinate when the inventory is managed by the retailer (RMI).In the third essay, we help in optimizing the pricing and ordering decisions of a retailer who sells a perishable product. We model this perishable product as having a shelf life of two periods. In the first period, the product is "new" product and in the next, it becomes "old" product. Every period, the retailer decides on what the prices for new and old products should be and how much of the new product to order. We model customer arrivals at the retailer by a Poisson process. Any customer has three choices buy a unit of old product, buy a unit of new product or not buy anything. We model customer preference among these choices using the multinomial logit (MNL) model our model also accounts for the potential dynamic demand substitution by customers between old and new products. We analyze and compare different control policies of the retailer by modeling the retailer's decision choice problem as a Markov decision problem under infinite horizon.