Approximate steady-state distribution for a large repairable item inventory system (Q1095509)

This paper examines a model for a repairable inventory system where some items are condemned upon failure. These condemnations require an ordering policy which we take to be an (s,S) policy. We model this system as a version of the classical machine repair problem using continuous time Markov processes. We develop solution procedures based on approximations to the steady state probability distribution, which enable us to obtain the operating characteristics of (s,S) policies very easily, with or without an ordering leadtime. The approximate models are found to be very accurate and computationally very efficient.