A Unifying Approach to Prioritisation and Optimisation of Maintenance
Here we present a unifying and integrating approach which can be used for long term optimisation, priority setting, and combination of maintenance. The approach focuses on the timing aspect of an action to be taken, like replacement, overhaul, or inspection. This is a typical aspect for many maintenance models. A penalty cost function is defined which expresses the expected costs of deferring the action to the next decision moment. The action is taken if the cost function exceeds a threshold level. We will show that several models fit within this structure, such as the block replacement model, the minimal repair model, an efficiency model, and also an opportunity block replacement model. The penalty cost function is also suitable for data elicitation purposes and avoids that questions are model dependent. This plays a role in setting up userinterfaces for decision support systems incorporating maintenance optimisation models. The penalty function can also serve as priority function, in case execution of activities has to be deferred because of capacity constraints. Furthermore, it is useful in combining activities if a first planning has been made by separate models (a decomposition approach). Finally, it can serve as a heuristic to derive structured maintenance policies, which we will demonstrate with a group replacement problem.