Hybridation de prouveurs CSP et apprentissage

Today, the MGAC-dom/wdeg algorithm, i.e., the algorithm which maintains Generalized Arc Consistency during the search of a solution, is still considered as the most efficient generic approach to cope with large and hard problem instances of the Constraint Satisfaction Problem (CSP). In this paper, an hybrid approach combining tree search with local search is proposed. Several combination schemes are considered, combining nogood learning, constraint weighting, Tabu lists, restarts and greedy assignments. On many CSP instances, experimental results of our approach show interesting improvements on the well known MGACdom/wdeg algorithm.