Multi-objective optimisation of asset maintenance management

Asset managers are faced with the necessity to reduce total life-cycle costs of their facilities as well as to improve their physical condition and long-term performance while considering the relative consequences of their failure. This paper presents a systematic decision-making approach for asset maintenance management that combines a stochastic Markovian performance prediction model with a multi-objective optimisation method to determine the optimal allocation of funds and prioritisation of assets for maintenance, repair and replacement. The maintenance optimisation is formulated as a multi-stage and multi-objective optimisation problem where the simultaneous satisfaction of several conflicting objectives, including minimisation of maintenance costs, maximisation of condition and performance, and minimisation of risk (weighted consequence) of failure, is sought. Compromise programming is used to determine the optimal ranking of deteriorated assets in terms of their priority for repair or replacement, by achieving a satisfactory trade-off between the competing / conflicting objectives. In addition, the paper examines the benefits of this approach as compared to the traditional maintenance prioritisation and management practices.