An infeasible solutions diversity maintenance epsilon constraint handling method for evolutionary constrained multiobjective optimization

It is well known that it is very difficult to solve constrained multiobjective optimization problems. Such problems not only need to optimize the objective function but also need to consider the constraints. The epsilon constraint handling method is commonly used, which releases the degree of constr...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Soft computing (Berlin, Germany) Ročník 25; číslo 13; s. 8051 - 8062
Hlavní autoři: Zhou, Jinlong, Zou, Juan, Zheng, Jinhua, Yang, Shengxiang, Gong, Dunwei, Pei, Tingrui
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.07.2021
Témata:
Artificial Intelligence
Computational Intelligence
Control
Engineering
Foundations
Mathematical Logic and Foundations
Mechatronics
Robotics
Constraints
Epsilon constraint handling method
Multiobjective optimization
Diversity
ISSN:1432-7643, 1433-7479
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:It is well known that it is very difficult to solve constrained multiobjective optimization problems. Such problems not only need to optimize the objective function but also need to consider the constraints. The epsilon constraint handling method is commonly used, which releases the degree of constraint violations by defining a gradually decayed epsilon. However, for the solutions whose overall constraint violations degree is greater than epsilon, the original epsilon constraint handling method cannot guarantee the diversity of solutions and only constraint violations are considered. To solve this issue, this paper proposed an infeasible solutions diversity maintenance strategy for solutions whose constraint violations degree is greater than epsilon. The experimental results show that our proposed algorithm is very competitive with other state-of-the-art algorithms for constrained multiobjective optimization problems.
ISSN:1432-7643
1433-7479
DOI:10.1007/s00500-021-05880-5