A dual-stage differential evolution algorithm based on diversity quality for multimodal multiobjective optimization

doi:10.1016/j.asoc.2026.115380

A dual-stage differential evolution algorithm based on diversity quality for multimodal multiobjective optimization

2026-05-08

0

PRE

AI

DOI：10.1016/j.asoc.2026.115380

Original

Original request for help

Save

Abstract

En 中文

• A dual-stage differential vector generation method is designed, in which the new solutions are reproduced in different niches and the new solutions are generated according to the distance among the whole population in the later stage. This method can effectively expand the search range of the population for finding more equvalent pareto optimal solutions. • For dominated-based MMOEAs, a new selection strategy is introduced, in which the diversity quality of each solution is calculated according to the distribution of the selected solutions in the decision space. • A dynamic preference deletion operation and an external archive of diversity are used to help the population to obtain a good distribution in both spaces. Experimental comparisons between TNLIDE and other competitive MMOEAs highlight the superior performance of TNLIDE on the CEC2019 benchmark test set.

Keywords:

dual-stage differential evolution

multimodal multiobjective optimization

diversity quality

Pareto optimal solutions

dynamic preference deletion

A dual-stage differential evolution algorithm based on diversity quality for multimodal multiobjective optimization

Abstract

Journal

Organization