Integration of maintenance scheduling and planning for large-scale asset fleets

Large fleets of engineering assets that are subject to ongoing degradation are posing the challenge of how and when to perform maintenance. For a given case study, this paper proposes a formulation for combined scheduling and planning of maintenance actions. A hierarchical approach and a two-stage approach (with either uniform or non-uniform time grid) are considered and compared to each other. The resulting discrete-time linear programming model follows the Resource Task Network framework. Asset deterioration is considered linearly and tackled with an enumerator-based formulation. Advantages of the model are its computational efficiency, scalability, extendability and adaptability. The results indicate that combined maintenance planning and scheduling can be solved in appropriate time and with appropriate accuracy. The decision-support that is delivered helps the choice of the specific maintenance action to perform and proposes when to conduct it. The paper makes a case for the benefits of optimally combining long-term planning and short-term scheduling in industrial-sized problems into one system.