Evolutionary Algorithm for Multiobjective Optimization Based on Density Estimation Ranking

Abstract

In the past few decades, a number of multiobjective evolutionary algorithms (MOEAs) have been proposed in the continue study. As pointed out in some recent studies, the performance of the most existing MOEAs is not promising when solving different shapes of Pareto fronts. To address this issue, this paper proposes an MOEA based on density estimation ranking. The algorithm includes density estimation ranking to shift the reference solution position, calculating the density of candidate solutions and ranking by the estimated density value, to modify the Pareto dominance relation and for handling complicated Pareto front. The result of this ranking can be used as the second selection criterion for environmental selection, and the optimal candidate individual with distribution and diversity information is selected. Experimental results show that the proposed algorithm can solve various types of Pareto fronts, outperformance several state-of-the-art evolutionary algorithms in multiobjective optimization.