MOEA/D with Adaptive Weight Vector Design

Abstract

MOEA/D (multi-objective evolutionary algorithm based on decomposition) has become a promising evolution algorithm for many-objective optimization problems, and many scholars conduct their researches on how to generate weight vectors in improved MOEA/D. In order to generate uniform non-dominated solution set according to the geometric shape of the Pareto front, an adaptive weight vector design method combining generalized decomposition and uniform design is proposed, which will adjust the setting the weight vector dynamically. The results for two standard test functions show that the proposed algorithm has certain advantages in convergence and diversity compared with other related algorithms, such as MOEA/D and UMOEA/D.