Simulation based optimization for maintenance strategies using Altarica 3.0
Résumé
The maintenance of industrial production systems is one of the major actual challenges, and the management of their maintenances is an important competitiveness factor. In fact a suitable maintenance strategy increases the system availability, and decreases costs due to interventions. Different kinds of maintenance policies can be used: corrective ones to repair the system after failures, preventive ones to maintain the system before failure. Based on a system case study, this work is a contribution to the field of simulation-based optimisation of maintenance strategies. The goal is to plan preventive maintenances according to two objectives: system availability and cost minimisation. Stochastic Discrete Event Simulation model was developed using Altarica 3.0 tool and connected to a Multi-objective optimisation algorithm (NSGAII). Both simulation and optimization modules interact with each other according to an iterative loop. The optimization combines statistics, coming from previous iteration, and the NSGAII algorithm provide to the simulation a new set of parameters for the model (the dates of maintenances), then the outputs of the simulation are performed in a new iteration. The experiments show that the NSGAII based approach achieved the best results although the computation time was relatively high and give more flexibility to the decision maker and give a trade-off between the two objectives.