A dual-population co-evolution algorithm with balanced environmental selection for constrained multimodal multiobjective optimization problems

In constrained multimodal multiobjective optimization problems (CMMOPs), the principal challenge is to explore multiple conflicting objectives and multiple equivalent Pareto sets under complex constraints, while balancing feasibility, convergence, and diversity of solutions. This paper proposes the...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Swarm and evolutionary computation Ročník 94; s. 101862
Hlavní autoři:	Wu, FuLong, Sun, Yu
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	Elsevier B.V 01.04.2025
Témata:	Balanced environmental selection Bi-crowding distance Constrained multimodal multiobjective optimization Dual-population co-evolution algorithm Dynamic-range-based constrained dominance principle Dual-population co-evolution algorithm Dynamic-range-based constrained dominance principle Balanced environmental selection Constrained multimodal multiobjective optimization Bi-crowding distance
ISSN:	2210-6502
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	In constrained multimodal multiobjective optimization problems (CMMOPs), the principal challenge is to explore multiple conflicting objectives and multiple equivalent Pareto sets under complex constraints, while balancing feasibility, convergence, and diversity of solutions. This paper proposes the DPCMMOEA-BES algorithm, which is based on dual-population co-evolution and incorporates a balanced environmental selection (BES) component to solve CMMOPs. In DPCMMOEA-BES, parent information from dual populations is shared through the mating selection operator based on speciation to generate offspring. Additionally, the BES component proposed in this paper enhances the algorithm’s overall performance by utilizing the dynamic-range-based constrained dominance principle and the accurate selection operation based on global Bi-crowding Distance, where the introduction of Bi-crowding Distance effectively balances the diversity of solutions in both the objective and decision spaces. The BES component also demonstrates its potential as a universal plugin, which can be integrated into various constrained multiobjective evolutionary algorithms and multimodal multiobjective evolutionary algorithms. The proposed DPCMMOEA-BES is evaluated on 31 test instances and compared with other state-of-the-art algorithms. The experimental results show that it is a highly competitive approach. Moreover, the comparative results confirm that integrating the BES component significantly improves the algorithm’s performance in solving CMMOPs.
ISSN:	2210-6502
DOI:	10.1016/j.swevo.2025.101862