Adaptive component crossover for differential evolution in solving single-objective optimization problems

With the increase of the complexity of engineering problems, evolutionary algorithms became an effective approach to black-box optimization problems. One of the most popular and promising evolutionary methods is the Differential Evolution algorithms. This method involves several evolutionary operato...

Full description

Saved in:
Bibliographic Details
Main Authors: Sopov Anton, Sopov Evgenii
Format: Article
Language:English
Published: EDP Sciences 2025-01-01
Series:ITM Web of Conferences
Online Access:https://www.itm-conferences.org/articles/itmconf/pdf/2025/03/itmconf_hmmocs-III2024_05001.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849717114755088384
author Sopov Anton
Sopov Evgenii
author_facet Sopov Anton
Sopov Evgenii
author_sort Sopov Anton
collection DOAJ
description With the increase of the complexity of engineering problems, evolutionary algorithms became an effective approach to black-box optimization problems. One of the most popular and promising evolutionary methods is the Differential Evolution algorithms. This method involves several evolutionary operators, including crossover, which is used to form offspring based on mutant and parent vectors, and is important in forming new generations of solutions. However, the classic differential evolution and its numerous modifications usually tends to use the single crossover mechanism to each of the variables of the system, therefore the properties and role of the subcomponents are not considered. That may lead to a slower convergence and increasing demands on computing resources. In this study we have proposed a novel Adaptive Component Crossover strategy for differential evolution, in which the crossover rate parameter is represented by a vector and its values are based on the behavior of the objective function on separate components. The experimental results on a set of benchmark problems have shown that the proposed scheme can improve the performance of the algorithm and, in particular, increase the convergence speed and crossover success rate.
format Article
id doaj-art-e6f0baf431024d3fac31e6d23e57534b
institution DOAJ
issn 2271-2097
language English
publishDate 2025-01-01
publisher EDP Sciences
record_format Article
series ITM Web of Conferences
spelling doaj-art-e6f0baf431024d3fac31e6d23e57534b2025-08-20T03:12:46ZengEDP SciencesITM Web of Conferences2271-20972025-01-01720500110.1051/itmconf/20257205001itmconf_hmmocs-III2024_05001Adaptive component crossover for differential evolution in solving single-objective optimization problemsSopov Anton0Sopov Evgenii1Reshetnev Siberian State University of Science and Technology, Institute of Informatics and TelecommunicationsReshetnev Siberian State University of Science and Technology, Institute of Informatics and TelecommunicationsWith the increase of the complexity of engineering problems, evolutionary algorithms became an effective approach to black-box optimization problems. One of the most popular and promising evolutionary methods is the Differential Evolution algorithms. This method involves several evolutionary operators, including crossover, which is used to form offspring based on mutant and parent vectors, and is important in forming new generations of solutions. However, the classic differential evolution and its numerous modifications usually tends to use the single crossover mechanism to each of the variables of the system, therefore the properties and role of the subcomponents are not considered. That may lead to a slower convergence and increasing demands on computing resources. In this study we have proposed a novel Adaptive Component Crossover strategy for differential evolution, in which the crossover rate parameter is represented by a vector and its values are based on the behavior of the objective function on separate components. The experimental results on a set of benchmark problems have shown that the proposed scheme can improve the performance of the algorithm and, in particular, increase the convergence speed and crossover success rate.https://www.itm-conferences.org/articles/itmconf/pdf/2025/03/itmconf_hmmocs-III2024_05001.pdf
spellingShingle Sopov Anton
Sopov Evgenii
Adaptive component crossover for differential evolution in solving single-objective optimization problems
ITM Web of Conferences
title Adaptive component crossover for differential evolution in solving single-objective optimization problems
title_full Adaptive component crossover for differential evolution in solving single-objective optimization problems
title_fullStr Adaptive component crossover for differential evolution in solving single-objective optimization problems
title_full_unstemmed Adaptive component crossover for differential evolution in solving single-objective optimization problems
title_short Adaptive component crossover for differential evolution in solving single-objective optimization problems
title_sort adaptive component crossover for differential evolution in solving single objective optimization problems
url https://www.itm-conferences.org/articles/itmconf/pdf/2025/03/itmconf_hmmocs-III2024_05001.pdf
work_keys_str_mv AT sopovanton adaptivecomponentcrossoverfordifferentialevolutioninsolvingsingleobjectiveoptimizationproblems
AT sopovevgenii adaptivecomponentcrossoverfordifferentialevolutioninsolvingsingleobjectiveoptimizationproblems