A Surrogate-Assisted Multi-objective Evolutionary Algorithm Guided by Hybrid Reference Points

Surrogate-assisted multi-objective evolutionary algorithms show excellent performance in solving expensive multi-objective optimization problems, but most of them do not work well for discontinuities Pareto fronts(PFs). To address this problem, a surrogate-assisted multi-objective evolutionary algorithm guided by hybrid reference points is proposed in this paper. The algorithm introduces a discontinuous region boundary point identification strategy to recognize the discontinuous information of PFs and set the reference points. Moreover, a two-stage multi-reference point-assisted management strategy is developed, which enables the algorithm to obtain better performance on different irregular discontinuous PFs. Experimental results show that the algorithm outperforms comparison algorithms on most of the problems with discontinuous PFs.