Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State

We present a novel approach to performing fitness approximation in genetic algorithms (GAs) using machine learning (ML) models, focusing on dynamic adaptation to the evolutionary state. We compare different methods for (1) switching between actual and approximate fitness, (2) sampling the population...

Full description

Saved in:
Bibliographic Details
Main Authors: Itai Tzruia, Tomer Halperin, Moshe Sipper, Achiya Elyasaf
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/15/12/744
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846104321401290752
author Itai Tzruia
Tomer Halperin
Moshe Sipper
Achiya Elyasaf
author_facet Itai Tzruia
Tomer Halperin
Moshe Sipper
Achiya Elyasaf
author_sort Itai Tzruia
collection DOAJ
description We present a novel approach to performing fitness approximation in genetic algorithms (GAs) using machine learning (ML) models, focusing on dynamic adaptation to the evolutionary state. We compare different methods for (1) switching between actual and approximate fitness, (2) sampling the population, and (3) weighting the samples. Experimental findings demonstrate significant improvement in evolutionary runtimes, with fitness scores that are either identical or slightly lower than those of the fully run GA—depending on the ratio of approximate-to-actual-fitness computation. Although we focus on evolutionary agents in Gymnasium (game) simulators—where fitness computation is costly—our approach is generic and can be easily applied to many different domains.
format Article
id doaj-art-e53931c3daaf4dd6929d114353edb711
institution Kabale University
issn 2078-2489
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Information
spelling doaj-art-e53931c3daaf4dd6929d114353edb7112024-12-27T14:30:42ZengMDPI AGInformation2078-24892024-11-01151274410.3390/info15120744Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary StateItai Tzruia0Tomer Halperin1Moshe Sipper2Achiya Elyasaf3Department of Computer Science, Ben-Gurion University, Beer-Sheva 8410501, IsraelDepartment of Computer Science, Ben-Gurion University, Beer-Sheva 8410501, IsraelDepartment of Computer Science, Ben-Gurion University, Beer-Sheva 8410501, IsraelDepartment of Software and Information Systems Engineering, Ben-Gurion University of the Negev, Beer-Sheva 8410501, IsraelWe present a novel approach to performing fitness approximation in genetic algorithms (GAs) using machine learning (ML) models, focusing on dynamic adaptation to the evolutionary state. We compare different methods for (1) switching between actual and approximate fitness, (2) sampling the population, and (3) weighting the samples. Experimental findings demonstrate significant improvement in evolutionary runtimes, with fitness scores that are either identical or slightly lower than those of the fully run GA—depending on the ratio of approximate-to-actual-fitness computation. Although we focus on evolutionary agents in Gymnasium (game) simulators—where fitness computation is costly—our approach is generic and can be easily applied to many different domains.https://www.mdpi.com/2078-2489/15/12/744genetic algorithmmachine learningfitness approximationsurrogate-assisted evolutionary algorithmregressionagent simulation
spellingShingle Itai Tzruia
Tomer Halperin
Moshe Sipper
Achiya Elyasaf
Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
Information
genetic algorithm
machine learning
fitness approximation
surrogate-assisted evolutionary algorithm
regression
agent simulation
title Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
title_full Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
title_fullStr Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
title_full_unstemmed Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
title_short Fitness Approximation Through Machine Learning with Dynamic Adaptation to the Evolutionary State
title_sort fitness approximation through machine learning with dynamic adaptation to the evolutionary state
topic genetic algorithm
machine learning
fitness approximation
surrogate-assisted evolutionary algorithm
regression
agent simulation
url https://www.mdpi.com/2078-2489/15/12/744
work_keys_str_mv AT itaitzruia fitnessapproximationthroughmachinelearningwithdynamicadaptationtotheevolutionarystate
AT tomerhalperin fitnessapproximationthroughmachinelearningwithdynamicadaptationtotheevolutionarystate
AT moshesipper fitnessapproximationthroughmachinelearningwithdynamicadaptationtotheevolutionarystate
AT achiyaelyasaf fitnessapproximationthroughmachinelearningwithdynamicadaptationtotheevolutionarystate