Two-Stage Greedy Approximated Hypervolume Subset Selection for Large-Scale Problems

Nan，Yang; Ishibuchi，Hisao; Shu，Tianye; Shang，Ke

10.1007/978-3-031-27250-9_28

2023

0302-9743

1611-3349

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

13970 LNCS

391-404

Recently, it has been demonstrated that a solution set that is better than the final population can be obtained by subset selection in some studies on evolutionary multi-objective optimization. The main challenge in this type of subset selection is how to efficiently handle a huge candidate solution set, especially when the hypervolume-based subset selection is used for many-objective optimization. In this paper, we propose an efficient two-stage greedy algorithm for hypervolume-based subset selection. In each iteration of the proposed greedy algorithm, a small number of promising candidate solutions are selected in the first stage using the rough hypervolume contribution approximation. In the second stage, a single solution among them is selected using the more precise approximation. Experimental results show that the proposed algorithm is much faster than state-of-the-art hypervolume-based greedy subset selection algorithms at the cost of a slight deterioration of the selected subset quality.

Evolutionary multi-objective optimization Hypervolume subset selection Two-stage hypervolume contribution approximation

First ; Corresponding

English

2-s2.0-85151065311

Scopus

Guangdong Provincial Key Laboratory of Brain-Inspired Intelligent Computation,Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,518055,China

Nan，Yang,Ishibuchi，Hisao,Shu，Tianye,et al. Two-Stage Greedy Approximated Hypervolume Subset Selection for Large-Scale Problems[C],2023:391-404.