An Improved Decomposition-based Multi-objective Evolutionary Algorithm with Enhanced Differential Evolution Strategy

The multi-objective evolutionary algorithm based on decomposition (MOEA/D) is proved to have a significant advantage dispose of multi-objective optimization problems (MOPs) since its introduction. However, initial MOEA/D worsen diversity, Furthermore, it is easy to generate an inferior solution by u...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2019 IEEE Symposium Series on Computational Intelligence (SSCI) s. 2245 - 2251
Hlavní autoři: Xie, Yingbo, Hou, Ying, Qiao, Junfei, Yin, Baocai
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.12.2019
Témata:
Convergence
Differential evolution
Elite archive
Evolutionary computation
Indexes
MOEA/D
Multi-objective optimization
Mutation strategy
Optimization
Performance evaluation
Sociology
Statistics
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The multi-objective evolutionary algorithm based on decomposition (MOEA/D) is proved to have a significant advantage dispose of multi-objective optimization problems (MOPs) since its introduction. However, initial MOEA/D worsen diversity, Furthermore, it is easy to generate an inferior solution by using simulated binary crossover operation. Therefore, an improved MOEA/D with an enhanced differential evolution (MOEA/D-EDE) is proposed to solve above problems. The newness and advantages of this proposed MOEA/D-EDE include the following two aspects. First, several differential evolution operators are used to replace crossover operator in the original MOEA/D. Second, MOEA/D-EDE introduces an elite archive strategy, thereby significantly increases the convergence speed while ensuring the diversity. Finally, the proposed MOEA/D-EDE is studied on MOPs compared with several MOEA/D variants and other algorithms. Empirical results display that MOEA/D-EDE to enhance the performance.
DOI:10.1109/SSCI44817.2019.9002905