Dynamic and adaptive grouping maintenance strategies: New scalable optimization algorithms

This paper focuses on new efficient and adaptive optimization algorithms to cope with the maintenance grouping problem for series, parallel, and complex systems. We propose a Particle Swarm Optimization approach to cope with small and medium problem sizes, and that will be used to benchmark existing heuristic solutions such as Genetic Algorithms. To address scalability and adaptability issues, we propose a new dynamic optimization algorithm based on a clustering technique. This clustering-based solution is formulated using an Integer Linear Programing approach to guarantee the convergence to global optimal solutions of the considered problem. We show the performance of the proposed approaches with a clear advantage to the clustering-based algorithm that we recommend for large industrial systems.