Core-guided method for constraint-based multi-objective combinatorial optimization

摘要: Multi-Objective Combinatorial Optimization(MOCO), which consists of several conflicting objectives to be optimized, finds an ever-increasing number uses in many real-world applications. In past years, the research MOCO mainly focuses on evolutionary algorithms. Recently, constraint-based methods come into view and have been proved effective problems. This paper builds previous works algorithm MCSEnumPD(AAAI-18) using path diversification method. Due inadequacy that original method fails prune redundant search space effectively, this proposes definition infeasible develops a novel exploits properties unsatisfiable cores, referred as CgPDMCS. The approach extends MCSEnumPD with core-guided method, avoids solving paths representing supersets cores. Experimental results show provides further performance gains over algorithms, especially for instances tightly constrained.