Semi-variance models in total productive maintenance
MetadataShow full item record
This research investigates the problem of finding an optimal time for performing preventive maintenance on a production system, under risk-sensitive semi-variance criterion. Risk-sensitive preventive maintenance policies, which use the variance of the cost as a measure of maintenance risk, have been widely studied in the literature. However, variance fails to differentiate between upside risk and downside risk. The purpose of this thesis is to demonstrate the superiority of using semi-variance to model the maintenance management risk. A new risk-sensitive framework is presented, with the objective of minimizing the long-run expected cost of performing preventive maintenance, and its semi-variance from a known target cost. The concept of Renewal theory and Markov chains, is used to develop mathematical models, which use the well-known Markowitz criterion, but with semi-variance as a measure of risk. We present computational results to analyze the performance of our models, and compare the results with risk-neutral and variance-penalized problem instances. Finally, an implementation of the Renewal theory model on a real-world problem is presented.