.png){kind=icon}
.png){kind=icon}
Return
Constrained multimodal multi-objective optimization based on cooperative evolution and neural network
.png){kind=icon}
PRE
AI
DOI：10.1007/s13042-025-02768-w.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Constrained multimodal multi-objective optimization problems (CMMOPs) require not only convergence, diversity and feasibility in the objective space, but also diversity in the decision space. However, when addressing CMMOPs, traditional multimodal multi-objective optimization evolutionary algorithms (MMOEAs) do not consider constraints, meanwhile, typical constrained multi-objective optimization evolutionary algorithms (CMOEAs) exhibit poor performance in solving multimodal problems. In these situations, this paper proposes a constrained multimodal multi-objective optimization evolutionary algorithm (CMMOEA) named CMMOCENN, which specializes in addressing CMMOPs. CMMOCENN utilizes a growing neural gas (GNG) network to enhance the convergence and diversity in the objective and decision space, respectively. In addition, a self-adaptive parameter adjustment considering objective functions is adopted to tackle constraints at different stages to satisfy the solution’s feasibility, further more, two environment selection strategies considering the diversity of decision space are devised for two different populations on evolution. Finally, the proposed CMMOCENN is verified on two test issues with several state-of-the-art algorithms. The experiment results demonstrate that CMMOCENN is competitive in solving CMMOPs.
Keywords:
Constrained multimodal multi-objective optimization
Evolutionary algorithm
Neural network
Self-adaptive parameter adjustment
Journal
IF:
2.7
Papers:
3.1K
Citations:
5.6K