Surrogate-assisted differential evolutionary algorithm with dynamic region exploration for expensive optimization problems

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27240%2F25%3A10260370" target="_blank" >RIV/61989100:27240/25:10260370 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.sciencedirect.com/science/article/pii/S1568494625009305?pes=vor&utm_source=clarivate&getft_integrator=clarivate" target="_blank" >https://www.sciencedirect.com/science/article/pii/S1568494625009305?pes=vor&utm_source=clarivate&getft_integrator=clarivate</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.asoc.2025.113619" target="_blank" >10.1016/j.asoc.2025.113619</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Surrogate-assisted differential evolutionary algorithm with dynamic region exploration for expensive optimization problems

Popis výsledku v původním jazyce

Currently, surrogate-assisted evolutionary algorithms (SAEAs) are widely used to solve computationally expensive optimization problems. However, in complex multimodal scenarios, there exists a notable likelihood that the algorithm might not converge to the optimal solution, highlighting the need to enhance the exploration capability of SAEAs. Therefore, exploring or developing different regions at different stages of the algorithm is crucial. This study designs a surrogate-assisted differential evolutionary algorithm with dynamic region exploration (DREDE) for expensive optimization problems. To filter out current suspected regions where better solutions might exist, anew criterion is proposed that combines the fitness level of a sample and its distance from the current optimal solution. By constructing surrogate models, DREDE dynamically performs global, local, and suspected region searches. A method for improving model accuracy by calculating the mean of selected individuals is introduced. These strategies synergistically enhance the performance of DREDE, which is comprehensively compared with several advanced SAEAs on seven benchmark functions with varying dimensions and was used in the reducer design problem. Simulation results showed that DREDE has a promising future in addressing costly practical issues.

Název v anglickém jazyce

Surrogate-assisted differential evolutionary algorithm with dynamic region exploration for expensive optimization problems

Popis výsledku anglicky

Currently, surrogate-assisted evolutionary algorithms (SAEAs) are widely used to solve computationally expensive optimization problems. However, in complex multimodal scenarios, there exists a notable likelihood that the algorithm might not converge to the optimal solution, highlighting the need to enhance the exploration capability of SAEAs. Therefore, exploring or developing different regions at different stages of the algorithm is crucial. This study designs a surrogate-assisted differential evolutionary algorithm with dynamic region exploration (DREDE) for expensive optimization problems. To filter out current suspected regions where better solutions might exist, anew criterion is proposed that combines the fitness level of a sample and its distance from the current optimal solution. By constructing surrogate models, DREDE dynamically performs global, local, and suspected region searches. A method for improving model accuracy by calculating the mean of selected individuals is introduced. These strategies synergistically enhance the performance of DREDE, which is comprehensively compared with several advanced SAEAs on seven benchmark functions with varying dimensions and was used in the reducer design problem. Simulation results showed that DREDE has a promising future in addressing costly practical issues.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10200 - Computer and information sciences

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Applied Soft Computing

ISSN

1568-4946

e-ISSN

1872-9681

Svazek periodika

183

Číslo periodika v rámci svazku

Nov

Stát vydavatele periodika

NL - Nizozemsko

Počet stran výsledku

17

Strana od-do

nestránkováno

Kód UT WoS článku

001541799100006

EID výsledku v databázi Scopus

—