Multi-objective differential evolution with dynamic hybrid constraint handling mechanism

Many real-world problems in engineering and process synthesis tend to be highly dimensional and nonlinear, even involve conflicting multiple objectives and subject to many constraints, which makes the feasible regions narrow; hence, it is hard to be solved by traditional constraint handling techniqu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Soft computing (Berlin, Germany) Jg. 23; H. 12; S. 4341 - 4355
Hauptverfasser: Lin, YueFeng, Du, Wei, Du, Wenli
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Berlin/Heidelberg Springer Berlin Heidelberg 01.06.2019
Springer Nature B.V
Schlagworte:
Artificial Intelligence
Computational Intelligence
Constraint modelling
Control
Engineering
Evolutionary algorithms
Evolutionary computation
Feasibility
Genetic algorithms
Handling
Mathematical Logic and Foundations
Mechatronics
Methodologies and Application
Methods
Multiple objective analysis
Mutation
Optimization
Robotics
Search strategies
Searching
Constraint handling
Differential evolutionary algorithm
Multi-objective optimization
Constrained optimization
ISSN:1432-7643, 1433-7479
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Many real-world problems in engineering and process synthesis tend to be highly dimensional and nonlinear, even involve conflicting multiple objectives and subject to many constraints, which makes the feasible regions narrow; hence, it is hard to be solved by traditional constraint handling techniques used in evolutionary algorithms. To handle this issue, this paper presents a multi-objective differential evolution with dynamic hybrid constraint handling mechanism (MODE-DCH) for tackling constrained multi-objective problems (CMOPs). In MODE-DCH, global search model and local search model combined with different constraint handling methods are proposed, and they are executed dynamically based on the feasibility proportion of the population. In the early stage that the feasible ratio is low, the local search model focuses on dragging the population into feasible regions rapidly, while the global search model is used to refine the whole population in the later stage. The two major modules of the algorithm cooperate together to balance the convergence and distribution of Pareto-optimal front. To demonstrate the effectiveness of MODE-DCH, the proposed algorithm is applied on several well-known CMOPs and two engineering problems compared with two other state-of-the-art multi-objective algorithms. The performance indicators show that MODE-DCH is an effective method to solve CMOPs.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1432-7643
1433-7479
DOI:10.1007/s00500-018-3087-z