A note on how we should express and interpret the results of stochastic maintenance optimization models Available to Purchase

Today most maintenance optimization models are established and interpreted within the classical statistical framework. There exist true failure time distributions and cost functions which we have to estimate. Using this approach, the results from the analysis are to a large extent disturbed by a discussion of uncertainty of the estimates. In this note we draw attention to an alternative approach: the fully Bayesian approach with focus on observable quantities and using subjective probabilities. We argue that this latter approach is more suitable as a tool for making decisions. The subjectivistic approach provides the framework of coherent use of (expert) judgment, which constitutes a significant (sometimes the only) part of information that is available to us.