The Role of Expert Judgement in Optimising Preventive Maintenance and System Architecture

Abstract-In the current industrial practice, optimisation of maintenance schedules is typically done using expert judgement but not via exhaustive exploration of all possible options for scheduling. Recently, it has been shown that search heuristics such as genetic algorithms can be used in conjunction with stochastic reliability prediction to optimise the maintenance schedules of components in a system. In this paper, we extend this framework to include the optional modelling of informed decisions by experts in terms of the time at which maintenance actions could be performed on components and decisions about which implementations of components should be in the final design. With this method, useful human knowledge and experience can be incorporated in a process that allows extensive exploration of the space of possible options for optimal or near optimal architecture and maintenance scheduling. The approach is demonstrated on a simplified model of fuel system. Keywords-maintenance; expert judgement; optimisation; genetic algorithm. I