An Analysis of Differential Evolution Population Size

The performance of the differential evolution algorithm (DE) is known to be highly sensitive to the values assigned to its control parameters. While numerous studies of the DE control parameters do exist, these studies have limitations, particularly in the context of setting the population size rega...

Full description

Saved in:
Bibliographic Details
Published in:Applied sciences 2024-11, Vol.14 (21), p.9976
Main Authors: Saad, Amani, Engelbrecht, Andries P., Khan, Salman A.
Format: Article
Language:English
Subjects:
Algorithms
Analysis
control parameters CP
Credibility
differential evolution DE
functional analysis of variance fANOVA
Genetic algorithms
Optimization
population size NP
Variance analysis
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The performance of the differential evolution algorithm (DE) is known to be highly sensitive to the values assigned to its control parameters. While numerous studies of the DE control parameters do exist, these studies have limitations, particularly in the context of setting the population size regardless of problem-specific characteristics. Moreover, the complex interrelationships between DE control parameters are frequently overlooked. This paper addresses these limitations by critically analyzing the existing guidelines for setting the population size in DE and assessing their efficacy for problems of various modalities. Moreover, the relative importance and interrelationship between DE control parameters using the functional analysis of variance (fANOVA) approach are investigated. The empirical analysis uses thirty problems of varying complexities from the IEEE Congress on Evolutionary Computation (CEC) 2014 benchmark suite. The results suggest that the conventional one-size-fits-all guidelines for setting DE population size possess the possibility of overestimating initial population sizes. The analysis further explores how varying population sizes impact DE performance across different fitness landscapes, highlighting important interactions between population size and other DE control parameters. This research lays the groundwork for subsequent research on thoughtful selection of optimal population sizes for DE algorithms, facilitating the development of more efficient adaptive DE strategies.
ISSN:2076-3417
2076-3417
DOI:10.3390/app14219976