A probabilistic pareto local search based on historical success counting for multiobjective optimization

In this paper, we propose a multiobjective probabilistic Pareto local search to address combinatorial optimization problems (COPs). The probability is determined by the success counts of local search offspring entering an external domination archive and this probabilistic information is used to further guide the selection of promising solutions for Pareto local search. In addition, simulated annealing is integrated in this framework as the local refinement process. This multiobjective probabilistic Pareto local search algorithm (MOPPLS), is tested on two famous COPs and compared with some well-known multiobjective evolutionary algorithms. Experimental results suggest that MOPPLS outperforms other compared algorithms.