| Productive or preventive maintenance (PM) is a widely-studied field in the manufacturing industry. Reducing system downtime and operating costs is one of the prime objectives of these policies. Renewal theory and Markov decision processes are one of the common theories used to model PM. Most of the literature focuses on the expected cost criterion only; the risk involved with each PM policy is not considered.;In this research, PM is studied via a renewal theory model and a Markov chain model. The concept of downside risk, which is employed in these models, is commonly used in financial portfolio management. The objective is to optimize the expected cost of operating a system subject to a constraint on the downside risk. The downside risk is modelled in this thesis as a probability of the cost per unit time exceeding a pre-set target. Computational results for both the models demonstrate a clear need for such models. Finally, this research could be further extended to other fields in stochastic systems research. |
