Multi-spatial information joint guidance evolutionary algorithm for dynamic multi-objective optimization with a changing number of objectives

Existing research on dynamic multi-objective optimization problems involving changes in the number of objectives has received little attention, but it is widespread in practical applications. This problem would cause the expansion or contraction of the manifold in the objective space. If it is accom...

Full description

Saved in:
Bibliographic Details
Published in:Neural computing & applications Vol. 35; no. 20; pp. 15167 - 15199
Main Authors: Ma, Xuemin, Sun, Hao, Hu, Ziyu, Wei, Lixin, Yang, Jingming
Format: Journal Article
Language:English
Published: London Springer London 01.07.2023
Springer Nature B.V
Subjects:
Artificial Intelligence
Computational Biology/Bioinformatics
Computational Science and Engineering
Computer Science
Data Mining and Knowledge Discovery
Dynamic response
Evolutionary algorithms
Genetic algorithms
Image Processing and Computer Vision
Multiple objective analysis
Original Article
Pareto optimization
Probability and Statistics in Computer Science
Spatial data
Multi-objective optimization
Dynamic multi-objective optimization
Evolutionary algorithm
Changing objectives
ISSN:0941-0643, 1433-3058
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Existing research on dynamic multi-objective optimization problems involving changes in the number of objectives has received little attention, but it is widespread in practical applications. This problem would cause the expansion or contraction of the manifold in the objective space. If it is accompanied by changes in Pareto set/front (PS/PF), the problem becomes more complex. However, several dynamic response techniques have been developed to handling this kind of dynamics. Faced with these issues, a multi-spatial information joint guidance evolutionary algorithm is proposed. To more accurately identify the optimal solutions after the change, a space adaptive transfer strategy is introduced, which adopts the geodesic flow kernel method to extract spatial information at different times. Afterwards it adaptively transfers the space via different changes to generate new individuals. In order to improve the diversity after the change, a dual space multi-dimensional joint sampling strategy is proposed. It fully combines the individual information in the objective and the decision space. Then the promising solutions are sampled in multiple dimensions near the representative individuals. Comprehensive experiments are conducted on 15 benchmark functions with a varying number of objectives and PS/PF. Simulation results verify the capability of the proposed algorithm.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0941-0643
1433-3058
DOI:10.1007/s00521-023-08369-4