Granular computing ranking method based multi-objective genetic algorithm

The key problem is the objective function sorting and fitness assignment in the Multi-Objective Evolutionary Algorithms(MOEAs). This paper regards the data generated in the process of the MOEAs as information system and introduces the method of the Granular Computing(GrC) to disposal the information system. Based on the dominate relationship in the information system, we get the dominance granule of the objective function, and adopt the granularity of dominance granule as the criteria of individual superiority, handle the incomparable characteristic of the Pareto solution set to form a quick sorting algorithm. Based on it, a multi-objective genetic algorithm is proposed. The result of the experiment shows that this method improves the efficiency of the MOEAs significantly and satisfies the requirements of the convergence.