Dynamic Rescheduling in the Context of Distributed Maintenance

Economic growth and global competition are pushing manufacturing companies to have several production sites, often relatively close to each other. We are interested in the maintenance management of geographically distributed production sites. This paper deals with the problem of unexpected failures that usually disrupt maintenance operations. The objective is to manage a fleet of homogeneous vehicles in charge of performing preventive and corrective maintenance operations at several production sites. The major difficulty lies in the number of sites to be considered, which makes the optimisation time exponential if the classical methods of the literature are used. Nevertheless, some studies allow to obtain an expected scheduling of maintenance operations considering the uncertain aspect of failures. The novelty of this paper is the development of an approach and an algorithm allowing to quickly and simply update the order of maintenance operations in case of unexpected events on line. Through an industrial case study in oil & gas field, we determine the optimal framework based on heuristics in which to modify the expected vehicle scheduling. The results show that with the appropriate parameters, simple rules can handle unexpected failures in a cost-effective way.