.png){kind=icon}
Return
Constrained subset-based two-stage evolutionary algorithm for constrained multi-objective optimization
.png){kind=icon}
PRE
AI
DOI：10.1016/j.engappai.2026.115582.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In constrained multi-objective optimization problems, when the number of constraints is large, it is difficult to effectively distinguish the influence of different types of constraints on the search process, thus reducing the search efficiency and convergence performance of the algorithm. To this end, this paper proposes the Constrained Subset-based Two-stage Evolutionary Algorithm for Constrained Multi-objective Optimization (CSTE) based on constraint subset. Firstly, the Constraint Subset Dynamic Grouping Strategy Based on Symbolic Collaboration Level is designed. By analyzing the collaborative relationship between constraints, the original constraint set is dynamically divided into multiple constraint subsets to realize the collaborative processing of different types of constraints. Secondly, the Dual-Domain Dual-Gradient-driven Auxiliary Population Search is proposed, which combines the feasibility gradient and the target gradient to guide the search direction of the auxiliary population, so as to improve the collaborative exploration ability of the algorithm to the feasible domain and the target space. Furthermore, the Elite Solutions Dynamic Migration Collaborative Optimization is constructed to guide the main population to gradually approach the real feasible Pareto frontier by dynamically migrating the high-quality solutions in the auxiliary population. Finally, the experimental results on 33 benchmark problems and three kinds of real engineering constrained multi-objective optimization scenarios show that the proposed algorithm has good comprehensive optimization ability in feasibility search, convergence performance and solution set distribution.
Journal
IF:
8
Papers:
5.7K
Citations:
3.5W
Organization
Cited Papers
.png){kind=icon}
ShareSave
ShareSaveShareSaveShareSave
View more