Using Dominated Solutions at Edges to the Diversity and the Uniformity of Non-dominated Solution Distributions in NSGA-II

Abstract This paper proposes a method for improving the diversity of Pareto front and uniformity non-dominated solution distributions in fast elitist sorting genetic algorithm (NSGA-II), which is an evolutionary multi-objective optimization algorithm. Conventional NSGA-II has excellent convergence to front, but it been reported that some test cases, does not produce more diverse distribution than strength 2 (SPEA2). In addition, selection using crowding distance may cause bias selected distribution. To avoid this problem, we propose archives dominated solutions be effective conventional search process when used operations, mates these archived with at edge rank 1 each objective function. We experimentally compare approach on typical ZDT suite problems two-objective constrained knapsack problem. By evaluating performance based diagrams, number solutions, maximum spread hypervolume values, show proposed both ends optimal