Self-adaptive parameters in differential evolution based on fitness performance with a perturbation strategy

Self-adaptive parameters in differential evolution based on fitness performance with a perturbation strategy

Differential evolution (DE) algorithms have been used widely to solve optimization problems and practical cases and have demonstrated high efficiency, performing favorably using only a few parameters. Compared with other traditional algorithms, DE algorithms perform well when used to solve continuou...

Full description

Saved in:
Bibliographic Details
Published in:Soft computing (Berlin, Germany) Germany), 2019-05, Vol.23 (9), p.3113-3128
Main Authors: Cheng, Chen-Yang, Li, Shu-Fen, Lin, Yu-Cheng
Format: Article
Language:eng
Subjects:
Adaptive algorithms
Algorithms
Artificial Intelligence
Computational Intelligence
Control
Engineering
Evolution
Evolutionary algorithms
Evolutionary computation
Fitness
Mathematical Logic and Foundations
Mechatronics
Methodologies and Application
Normal distribution
Optimization
Parameter identification
Parameters
Perturbation
Robotics
Traveling salesman problem
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Differential evolution (DE) algorithms have been used widely to solve optimization problems and practical cases and have demonstrated high efficiency, performing favorably using only a few parameters. Compared with other traditional algorithms, DE algorithms perform well when used to solve continuous problems. To obtain an approximate solution using DE, it is critical that appropriate parameter values are selected. However, selecting and dynamically tuning the parameter values during evolution are not easy tasks because the values depend significantly on the problem to be solved. To address these issues, this study presents an enhanced DE algorithm with self-adaptive adjustable parameters and a perturbation strategy based on individual fitness performance. Compared with two existing DE algorithms, the proposed algorithm can solve six benchmark functions and has both high efficiency and stability.
ISSN:1432-7643
1433-7479