Optimal (s,S) replenishment policies are determined for inventories of recoverable items in which an item occasionally is lost or must be discarded. Two types of recoverable items are considered: rental items and repairable spares. The objective is to minimize total expected order, holding, repair and shortage costs per unit of time. A discrete time Markov chain yields steady-state probabilities for the number of items in the system. Total cost functions are developed for two types of resupply time: immediate and deterministic. Examples are included to illustrate the sensitivity of the optimal (s,S) policies to parameter changes.
