A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization

Abstract Surrogate-assisted evolutionary algorithms (SAEAs) commonly depend on traditional offspring generation methods such as simulated binary crossover and polynomial mutation, which often lead to suboptimal search efficiencies. This paper introduces the gradient-descent-like learning-based SAEA...

Full description

Saved in:
Bibliographic Details
Main Authors: Chaoyi Sun, Bo Zhang, Hai Sun, Rui Feng
Format: Article
Language:English
Published: Springer 2025-06-01
Series:Complex & Intelligent Systems
Subjects:
Online Access:https://doi.org/10.1007/s40747-025-01955-0
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849331395245113344
author Chaoyi Sun
Bo Zhang
Hai Sun
Rui Feng
author_facet Chaoyi Sun
Bo Zhang
Hai Sun
Rui Feng
author_sort Chaoyi Sun
collection DOAJ
description Abstract Surrogate-assisted evolutionary algorithms (SAEAs) commonly depend on traditional offspring generation methods such as simulated binary crossover and polynomial mutation, which often lead to suboptimal search efficiencies. This paper introduces the gradient-descent-like learning-based SAEA (GDL-SAEA) designed for expensive multiobjective optimization problems. Our method aims to determine the learned possibly fastest convergence (i.e., gradient-descent-like) direction for each solution via a trained neural network, leveraging prior knowledge from the relationships within the current population and surrogate model. The population is segmented into three subgroups, generating triple training data through cosine similarity and the Mahalanobis distance. Notably, each elite solution within these groups serves as a label for the corresponding poor solution with a midpoint acting as an anchor, thereby enhancing the supervised learning of the convergence process. To implement the proposed framework, three representative SAEAs are embedded into the GDL-SAEA. Experimental evaluations on multiobjective benchmarks and real-world problems with up to 10 objectives reveal that GDL-SAEA outperforms seven state-of-the-art and classic algorithms.
format Article
id doaj-art-7741cd0ac1f44aa0aed1d01b0e85ef4f
institution Kabale University
issn 2199-4536
2198-6053
language English
publishDate 2025-06-01
publisher Springer
record_format Article
series Complex & Intelligent Systems
spelling doaj-art-7741cd0ac1f44aa0aed1d01b0e85ef4f2025-08-20T03:46:37ZengSpringerComplex & Intelligent Systems2199-45362198-60532025-06-0111812810.1007/s40747-025-01955-0A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimizationChaoyi Sun0Bo Zhang1Hai Sun2Rui Feng3College of Computer Science and Artificial Intelligence, Shanghai Key Laboratory of Intelligent Information Processing, Fudan UniversityShanghai Publishing and Printing CollegeSchool of Management, Fudan UniversityCollege of Computer Science and Artificial Intelligence, Shanghai Key Laboratory of Intelligent Information Processing, Fudan UniversityAbstract Surrogate-assisted evolutionary algorithms (SAEAs) commonly depend on traditional offspring generation methods such as simulated binary crossover and polynomial mutation, which often lead to suboptimal search efficiencies. This paper introduces the gradient-descent-like learning-based SAEA (GDL-SAEA) designed for expensive multiobjective optimization problems. Our method aims to determine the learned possibly fastest convergence (i.e., gradient-descent-like) direction for each solution via a trained neural network, leveraging prior knowledge from the relationships within the current population and surrogate model. The population is segmented into three subgroups, generating triple training data through cosine similarity and the Mahalanobis distance. Notably, each elite solution within these groups serves as a label for the corresponding poor solution with a midpoint acting as an anchor, thereby enhancing the supervised learning of the convergence process. To implement the proposed framework, three representative SAEAs are embedded into the GDL-SAEA. Experimental evaluations on multiobjective benchmarks and real-world problems with up to 10 objectives reveal that GDL-SAEA outperforms seven state-of-the-art and classic algorithms.https://doi.org/10.1007/s40747-025-01955-0Expensive multiobjective optimizationGradient-descent-like learningOffspring generationSurrogate-assisted evolutionary algorithmsNeural networks
spellingShingle Chaoyi Sun
Bo Zhang
Hai Sun
Rui Feng
A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
Complex & Intelligent Systems
Expensive multiobjective optimization
Gradient-descent-like learning
Offspring generation
Surrogate-assisted evolutionary algorithms
Neural networks
title A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
title_full A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
title_fullStr A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
title_full_unstemmed A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
title_short A gradient-descent-like learning-based framework in surrogate-assisted evolutionary algorithms for expensive many-objective optimization
title_sort gradient descent like learning based framework in surrogate assisted evolutionary algorithms for expensive many objective optimization
topic Expensive multiobjective optimization
Gradient-descent-like learning
Offspring generation
Surrogate-assisted evolutionary algorithms
Neural networks
url https://doi.org/10.1007/s40747-025-01955-0
work_keys_str_mv AT chaoyisun agradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT bozhang agradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT haisun agradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT ruifeng agradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT chaoyisun gradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT bozhang gradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT haisun gradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization
AT ruifeng gradientdescentlikelearningbasedframeworkinsurrogateassistedevolutionaryalgorithmsforexpensivemanyobjectiveoptimization