Optimal set-up of industrial scheduling software

For a long time, workshop managers have only expected from industrial schedulers to find quickly and sometimes interactively a feasible schedule. However, since the schedule governs the manufacturing resource usage, companies now require to optimize their schedule according to more or less complex criteria. This optimization is not only necessary in order to allow the company to remain competitive, but also in order to allow it to efficiently participate to Supply Chains in a Collaborative context. Many parameters are present in nowadays industrial schedulers which have a direct influence on the schedule quality. Tuning these parameters is a long and complex task if performed manually, which can be addressed by combinatorial optimization methods. Meta-heuristics have also been successfully used on specific scheduling problems, but the size of industrial problems and the necessity to provide a generic solving framework set specific difficulties. We show in this communication how several meta-heuristics can be integrated and combined in an environment allowing both to define a complex objective function, on the base of several elemental criteria, and to optimize the schedule by finding the best tuning of several tens of scheduling parameters, Typical results based on the cases of real workshops are provided.