Improving Local Search Algorithms via Probabilistic Configuration Checking

Configuration checking (CC) has been confirmed to alleviate the cycling problem in local search for combinatorial optimization problems (COPs). When using CC heuristics graph problems, a critical concept is configuration of vertices. All existing variants employ either 1- or 2-level neighborhoods vertex as its configuration. Inspired by idea that with different levels should have contributions solving COPs, we propose probabilistic (PC), which introduces probabilities at consider impact on strategy. Based PC, first (PCC), can be developed an automated and lightweight favor. We then apply PCC two classic COPs shown achieve good results CC, our preliminary confirm improves algorithms because alleviates problem.