A multi-echelon spare parts inventory system with emergency lateral shipments subject to a response time constraint

Mainframe manufacturers face intense competition both from manufacturers of similar technologies as well as from rival technologies, such as PCs and powerful Unix servers. At the same time, declining prices for high-tech products are squeezing out margins. Thus, success in this highly competitive environment requires not only producing the best product at the lowest price, but also providing an excellent customer service. Customer service includes both highly reliable mainframes as well as prompt support service. Indeed, excellent service is the key to maintaining loyal customers who will return for more business and who will also provide positive testimony for potential customers. Of course, high service level typically implies high inventory levels; unfortunately, in this industry a manufacturer's inventory investment in the system can be a significant fraction of total cost. Therefore, in this highly competitive market where the profit margins were being squeezed out, it is essential to cut costs, including inventory holding costs, as much as possible.; Thus, the focus of this dissertation is on the analysis of a multi-echelon spare parts inventory system of a mainframe manufacturer. The model we construct and analyze is characterized by hundreds of parts and customers, very low part failure rates, tight service level requirement, and many field depots. Also, we allow emergency lateral shipments. Our goal is to identify stocking policies for this system to minimize system-wide inventory holding cost.; Specifically, we consider three problems in this dissertation. We first consider a single field depot that stocks many parts, serves many customers, and is subject to a tight response time requirement. Our objective is to determine a near optimal inventory policy at the depot by determining the appropriate stocking levels for each part. We assume the leadtime between an outside supplier and the depot is deterministic. We develop two heuristics to solve the problem, one of which has a worst case bound of 100%. We also conduct an extensive numerical study and observe that the heuristics perform very well on real-world instances.; Next, we analyze the multi-echelon setting, where we consider many field depots and a single warehouse that replenishes stock at these depots. The warehouse also stores operational spare parts and acts as a repair facility. No lateral shipments among the depots are allowed. Our task is to determine the stocking levels at the depots and the warehouse. We develop a strategy that incorporates the heuristics developed for the single depot problem. An empirical study allows us to show that this heuristic is very effective for the two-echelon problem.; Finally, we incorporate emergency lateral shipments into the analysis of the two-echelon problem. We assume that we are given pooling groups, in which depots neighboring each other share their stock. We develop approximations to estimate the fraction of demand satisfied by stock on hand, by lateral shipment, or backordered for each depot. These approximations are used in the development of a heuristic to solve the general problem. We evaluate the effectiveness of the heuristic against a lower bound using a computational study.