Differential evolution using multi-strategy for the improvement of optimization performance

Differential evolution (DE) is an effective population-based optimization approach that has been widely used to deal with scientific and engineering problems. However, the performance of DE method is largely dependent on its trial vector produce strategy, namely, mutation strategy, crossover operati...

Full description

Saved in:
Bibliographic Details
Published in:Neural computing & applications Vol. 37; no. 27; pp. 22593 - 22620
Main Authors: Liu, Nengxian, Luo, Jianbin, Chang, Jie, Pan, Jeng-Shyang
Format: Journal Article
Language:English
Published: London Springer London 01.09.2025
Springer Nature B.V
Subjects:
Adaptation
Archives & records
Artificial Intelligence
Computational Biology/Bioinformatics
Computational Science and Engineering
Computer Science
Data Mining and Knowledge Discovery
Evolutionary computation
Exploitation
Image Processing and Computer Vision
Methods
Mutation
Normal distribution
Optimization
Optimization algorithms
Parameters
Probability and Statistics in Computer Science
S.I.: Hybrid Approaches to Nature-inspired Optimization Algorithms and Their Applications
Special Issue on Hybrid Approaches to Nature-inspired Optimization Algorithms and Their Applications
Stagnation
Parameter adaptation schemes
Local search
Multi-strategy
Stagnation problem
ISSN:0941-0643, 1433-3058
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Differential evolution (DE) is an effective population-based optimization approach that has been widely used to deal with scientific and engineering problems. However, the performance of DE method is largely dependent on its trial vector produce strategy, namely, mutation strategy, crossover operation and its corresponding control parameters. As claimed by the ‘No free Lunch theorem’, each mutation or crossover strategy has its fatal flaws; hence, the DE method having a single operation strategy cannot solve all types of optimization problems. Therefore, we propose a novel multi-strategy DE (MS-DE) in this study. First, the proposed algorithm uses combined mutation strategies including two powerful mutation strategies and selects them in a probabilistic way. Second, an improved crossover operation is introduced to tackle the stagnation problem. When a stagnation occurs, DE employs the top p-best vector to conduct crossover operation. Third, the control parameters are tuned in novel adaptation schemes. Finally, a local search is utilized in the proposed method to accelerate the convergence. The proposed MS-DE method is examined on CEC2017 test suite, and experiment results confirm its outperformance over several state-of-the-art DE methods. Furthermore, the proposed MS-DE is applied to two constrained engineering problems. The comparison results on these two problems also demonstrate the efficiency of our proposed MS-DE.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0941-0643
1433-3058
DOI:10.1007/s00521-024-10781-3