A Fast Elitist Non-dominated Sorting Genetic Algorithm for Multi-objective Optimisation: NSGA-II
PPSN VI Proceedings of the 6th International Conference on Parallel Problem Solving from Nature
Simulation optimization: simulation-based optimization
Proceedings of the 34th conference on Winter simulation: exploring new frontiers
| Hi-index | 0.00 |
Properly executed maintenance is a key enabler for the availability of military systems. Hence, increasingly complex military and other technical systems require increasingly sophisticated support solutions. Condition-based maintenance (CBM) is applied to reduce downtime for scheduled maintenance while trying to minimize the need for unscheduled maintenance. A prerequisite for successful application of CBM is condition monitoring, with a corresponding architecture to manage the resulting volumes of data. A maintenance support solution has been developed at BAE Systems Bofors AB, integrating information and communication throughout the maintenance support solution infrastructure. At the lowest level, the architecture involves sensors on each system to track wear and usage. The highest level of the architecture consists of databases with configuration management, usage history, life-cycle models, etc. Information from sensors, databases and simulations needs to be integrated and fused to generate high-level decision support for maintenance planning. This paper outlines the maintenance support framework developed by BAE Systems Bofors AB and examines a novel approach of providing decision support for high-level maintenance planning. A genetic algorithm for simulation-based multi-objective optimization is used to examine the possible planning options. A Pareto-optimal set of plans is found, providing the planner with an overview of the best options available, as well as the means to make an informed trade-off analysis.