A novel strategy to preserve diversity in solving MOEAs based on artificial emphasis on uniformity

Diversity distribution plays a significant role in assessing performances of MOEAs, so we proposed LRCD (Left-Right Crowding Distance) in to get a better performance than that of the original NSGA2. There are two aspects of diversity, they are span and uniformity. Based on LRCD, we will propose W-LRCD and E-LRCD to put artificial emphasis on uniformity in calculation of crowding distance in this paper. Six Multi-Objective test problems have been done to test the two methods´ performances, we find E-LRCD performs the best, LRCD the second, W-LRCD the third and the original NSGA2 ranked the last on diversity distribution in the six test problems.