Embedding decision-maker's preferences in the multi-objective Tabu search method for scheduling problems
Résumé
The integration of the decision maker's preferences in the optimization process of a multi-objective problem is a currently prominent research topic. In this work, we consider the scheduling problem of a flexible shop with multiple objectives, where the makespan, the total machining time and the maximum workload all have to be minimized. We assume that the decision maker can express multiple aspiration levels for each of the three objectives, as well as additional information that allows us to construct a preference model. This model corresponds to a SRMP (Simple Ranking using Multiple Profiles) model which we propose to integrate in a new Tabu search based approach. The SRMP model serves to order the neighboring solutions and thus help in implementing the descent phase, while the aspiration levels are also used to filter out neighboring solutions without needing to decode them beforehand. The proposal is compared to a state-of-the-art Tabu search algorithm on benchmark instances. Preliminary results show that our approach is quicker to find better solutions from the perspective of the decision maker.