Maintenance Optimisation with Imperfect Information
Condition-based maintenance is an approach that recommends taking maintenance decisions based on measurements of the condition of a system. Ideally, decisions are based on full knowledge of a system’s condition and a completely specified (stochastic) deterioration model to describe how its condition will evolve in the future. However, in reality often one or more quantities of interest are only partially known. In this talk, we will discuss maintenance optimization problems that arise in settings with different types of imperfect information.
In the main part of the talk, we focus on the problem of optimally scheduling replacements for a Markovian deteriorating system under population heterogeneity. Heterogeneities in the population of components can arise, for example, because of variations in the production process of components. We assume that the population of spare components is composed of multiple component types that cannot be distinguished by their exterior appearance, but deteriorate according to different transition probability matrices. We formulate the problem using a partially observable Markov decision process (POMDP) model, and provide a set of conditions for which we characterize the structure of the optimal policy that minimizes the total expected discounted operating and replacement cost over an infinite horizon. By a numerical experiment, we benchmark the optimal policy against a heuristic policy that neglects population heterogeneity.
- Registration to Remy Spliet, firstname.lastname@example.org is required for availability of lunch.