Enhanced decomposition-based hybrid evolutionary and gradient-based algorithm for many-objective optimization

This paper presents a novel decomposition-based hybrid many-objective optimization method using particle swarm optimization (PSO) and sequential quadratic programming (SQP) algorithms. In the proposed method, the objective function space decomposed into optimization sub-problems. In this respect, th...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Applied intelligence (Dordrecht, Netherlands) Jg. 53; H. 24; S. 30497 - 30522
Hauptverfasser: Mohammad Zadeh, Parviz, Mohagheghi, Mostafa
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.12.2023
Springer Nature B.V
Schlagworte:
Artificial Intelligence
Computational efficiency
Computer Science
Coordinates
Decomposition
Evolutionary algorithms
Function space
Machines
Manufacturing
Mechanical Engineering
Multiple objective analysis
Operators (mathematics)
Optimization algorithms
Particle swarm optimization
Processes
Quadratic programming
Sequential quadratic programming (SQP)
Multi-objective particle swarm optimization (MOPSO)
Decomposition
Many-objective optimization
Reference point
Multi-objective evolutionary algorithms (MOEAs)
ISSN:0924-669X, 1573-7497
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper presents a novel decomposition-based hybrid many-objective optimization method using particle swarm optimization (PSO) and sequential quadratic programming (SQP) algorithms. In the proposed method, the objective function space decomposed into optimization sub-problems. In this respect, the origin of the coordinate systems of the scaled objective function space (i.e., ideal point) and uniformly distributed points in the scaled objective function space are used to generate reference directions. The scaling of the objective function space in each generation leads to the improvement of the accuracy of the ideal point, which is important because of its role in the generation of the reference directions. In the proposed method, optimization of all the reference directions performed simultaneously to increase the accuracy as well as the computational efficiency. In addition, the proposed method is based on the hybridization of multi-objective particle swarm optimization (MOPSO) using the reference direction and local fast and accurate search capabilities of the SQP to provide a computationally more efficient and accurate algorithm for solving many-objective optimization problems. In the proposed method, a new mutation operator was introduced into the MOPSO to enhance the computational performance, uniformity of the solution and to prevent the MOPSO from pre-convergence to the local Pareto front. The proposed method was compared with other many-objective optimization algorithms using several challenging DTLZ benchmark problems and real-world many-objective optimization problems involving 3 to 15 objective functions. The results show that the proposed method has a high diversity and very accurate Pareto solutions. In this respect, the proposed method outperforms other recently developed many-objective optimization algorithms used in this study. Based on the obtained results, the proposed method provides an effective way of solving many-objective optimization problems.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0924-669X
1573-7497
DOI:10.1007/s10489-023-05106-1