A Path-Repair Algorithm for Solving Scheduling Problems

Search algorithms for solving CSP (Constraint Satisfaction Problems) usually fall into one of two main families: local search algorithms and systematic algorithms. Both families have their own advantages, Designing hybrid approaches seem promising since those advantages may be combined into a single approach. In this paper, we propose a new hybrid technique which performs an overall local search over partial assignments instead of complete assignments, and uses filtering and conflict-based techniques to efficiently guide the search and maintain local consistency. In the case of inconsistency, it proposes explanations to retract only those constraints that are really involved in the contradiction. This new technique benefits from both classical approaches: a priori pruning of the search from filtering-based search and possible repair of early mistakes from local search. We focus on a specific version of local search technique called best-first-repair which determines, upon the constraint inconsistency, a pertinent choice point to repair the search path to restore the satisfiability. Experiments done on open-shop scheduling problems have shown with the best highly specialized algorithms.