Operating characteristics of a two-echelon inventory system for repairable items with condemnation allowed

Abstract

In this thesis, operating characteristics of a two-echelon inventory system for repairable items with condemnation allowed are considered. The upper echelon corresponds to a central depot, and the lower echelon consists of several bases. Each location in the corresponding echelon, representing an item stocking point, has repair facilities. The demands for a particular item at bases occur in Poison processes. Spares of the item can be stocked at base or depot. If a failed item can be repaired at a base, a replacement item will be supplied from the base spare stock, and the failed item is put in the spare stock after being repaired. Spares stocked at the base are issued on a FCFS priority and all demands are backordered until filled. If the item can not be repaired at the base, a spare will still be supplied from the base stock and a replenishment spare then be requested from the depot. The demands at the depot are also filled on a FCFS priority and backordered until filled. Items repaired at the depot are put into its spare stock. All the replenishment at the depot can be made from an external supplier with infinite items. Repair may be accomplished at the full capacity of all facilities. However, if any item is found not economical to repair, then it will be condemned. Such condemnation is known at a given rate in total. The thesis considers three structurally-different cases. In the first case, base repair process is regarded as a queueing system with a single repair channel. In the system, multiple types of failed items are repaired. however, infinite repair channels are assumed for the depot repair process. the depot is operated in a continuous review (s,S) inventory policy, but the bases are managed in one-for-one replenishment policies. General distributions are allowed for the repair time distributions of all facilities. In addition to the operating characteristics of the inventory system, an approximation method for computing probability distribution of...