An Analysis of a Single Location Inventory Problem for Two Interchangeable Recoverable Items.
Abstract
In this paper we examine the interchangeability/substitutability problem for two recoverable items that fail at a single location. We assume the failure processes for each type of item are independent, stationary Poisson processes. We also assume the repair times are exponentially distributed. Furthermore, we assume that the system is a closed system, that is, no items are added to or deleted from the system. We first consider a discrete-time problem and show that this problem is a Markovian decision problem. We then show that for this problem there exist optimal stationary Markov control policies. Next we formulate a continuous time model and show how to find the optimal stationary Markov control policy using linear programming. Unfortunately, this approach is impractical for solving most real problems. Consequently we have established and explored some of the properties that we feel an optimal policy should possess. A discussion of these properties is given in Section IV. Lastly, we will describe a heuristic that can be used to find a good policy. This method is an efficient simulation search method that finds policies having the properties we conjecture an optimal policy should possess. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1979
- Accession Number
- ADA067591
Entities
People
- Carol Shilepsky
- David Heath
- John Muckstadt
Organizations
- Cornell University College of Engineering