Dynamic Inventory Policies for Short Lifecycle and Perishable Products with Demand Uncertainty

This dissertation considers a common thread of inventory modeling for short lifecycle and perishable products. These nontraditional products introduce additional elements of complexity, particularly when the demand processes for these goods are nonstationary. We motivate our research with real world context from the consumer electronics and healthcare industries.;In Chapter 2, we consider a firm with a geographically distant supplier that simultaneously decides the order quantity for each of two supply modes (e.g., air shipment and ocean shipment) on a periodic basis. In each period, the firm may order from both sources, order from one source, or order nothing. This research extends the literature on two modes of supply by deriving the optimal solutions under nonstationary product lifecycle (PLC) dependent demand, with lead time differences between the two modes of one or more periods, and a finite sales horizon. This fluctuating demand environment is a distinctive feature of our model since both the demand mean and demand variability may change over time representing the PLC. We develop a Markov decision process (MDP) which determines the optimal ordering quantities per mode under stochastic, nonstationary demand. We consider two cases: Case 1 models demand evolution as an embedded Markov chain, and Case 2 considers the special case where demand evolution is time-dependent. Optimal base stock policies, defined by two base stock levels, are derived and shown to be both state and time-dependent for Case 2. Our results suggest that optimal shipping strategies (i.e., the distribution of total order quantities between each mode of supply) change throughout the PLC as demand characteristics change. We derive insights from the optimal structure of the time-dependent demand case for the general model. Further, we use numerical examples to explore the relationship between the optimal solutions and the phases of the PLC.;In Chapters 3 and 4, we model and analyze production and inventory policies for perishable pharmaceuticals in the context of the hospital pharmacy. Hospital pharmacies throughout the United States are experiencing drug inventory problems that result in waste and shortages that affect patient outcomes due to delayed procedures and drug substitutions. We consider a pharmacist's decision to order and hold drug products in two inventory stages (i.e., raw material and dispensed form) with varying shelf lives.;In Chapter 3, our research objective is to determine the optimal two-stage inventory and production policy. Since demand for medications is dependent on the patient mix, we define a stochastically changing 'demand state' as a surrogate for patient condition information that is represented by a Markov chain. We evaluate two demand fulfillment scenarios when a shortage occurs which depends on production capabilities. We formulate this problem as a Markov decision problem (MDP) and prove the existence of optimal solutions for both scenarios. Finally, we present numerical examples which demonstrate the behavior of the optimal solutions for various cost structures. We apply this model in Chapter 4 using demand data from a large public hospital. We develop a Markovian demand process and solve for the optimal raw material inventory order and finished good production quantities.