Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning

This paper investigates the application of genetic algorithms (GAs) for hyperparameter optimisation in deep reinforcement learning (RL), focusing on the Deep Q-Learning (DQN) algorithm. This study aims to identify approaches that enhance RL model performance through the effective exploration of the...

Full description

Saved in:
Bibliographic Details
Main Authors: Bartłomiej Brzęk, Barbara Probierz, Jan Kozak
Format: Article
Language:English
Published: MDPI AG 2025-02-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/4/2067
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849718864933289984
author Bartłomiej Brzęk
Barbara Probierz
Jan Kozak
author_facet Bartłomiej Brzęk
Barbara Probierz
Jan Kozak
author_sort Bartłomiej Brzęk
collection DOAJ
description This paper investigates the application of genetic algorithms (GAs) for hyperparameter optimisation in deep reinforcement learning (RL), focusing on the Deep Q-Learning (DQN) algorithm. This study aims to identify approaches that enhance RL model performance through the effective exploration of the configuration space. By comparing different GA methods for selection, crossover, and mutation, this study focuses on deep RL models. The results indicate that GA techniques emphasising the exploration of the configuration space yield significant improvements in optimisation efficiency, reducing training time and enhancing convergence. The most effective GA improved the fitness function value from 68.26 (initial best chromosome) to 979.16 after 200 iterations, demonstrating the efficacy of the proposed approach. Furthermore, variations in specific hyperparameters, such as learning rate, gamma, and update frequency, were shown to substantially affect the DQN model’s learning ability. These findings suggest that exploration-driven GA strategies outperform GA approaches with limited exploration, underscoring the critical role of selection and crossover methods in enhancing DQN model efficiency and performance. Moreover, a mini case study on the CartPole environment revealed that even a 5% sensor dropout impaired the performance of a GA-optimised RL agent, while a 20% dropout almost entirely halted improvements.
format Article
id doaj-art-39c0ddea08df4adb9d4b1fa749e5d63b
institution DOAJ
issn 2076-3417
language English
publishDate 2025-02-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-39c0ddea08df4adb9d4b1fa749e5d63b2025-08-20T03:12:16ZengMDPI AGApplied Sciences2076-34172025-02-01154206710.3390/app15042067Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement LearningBartłomiej Brzęk0Barbara Probierz1Jan Kozak2Department of Machine Learning, University of Economics in Katowice, 1 Maja 50, 40-287 Katowice, PolandDepartment of Machine Learning, University of Economics in Katowice, 1 Maja 50, 40-287 Katowice, PolandDepartment of Machine Learning, University of Economics in Katowice, 1 Maja 50, 40-287 Katowice, PolandThis paper investigates the application of genetic algorithms (GAs) for hyperparameter optimisation in deep reinforcement learning (RL), focusing on the Deep Q-Learning (DQN) algorithm. This study aims to identify approaches that enhance RL model performance through the effective exploration of the configuration space. By comparing different GA methods for selection, crossover, and mutation, this study focuses on deep RL models. The results indicate that GA techniques emphasising the exploration of the configuration space yield significant improvements in optimisation efficiency, reducing training time and enhancing convergence. The most effective GA improved the fitness function value from 68.26 (initial best chromosome) to 979.16 after 200 iterations, demonstrating the efficacy of the proposed approach. Furthermore, variations in specific hyperparameters, such as learning rate, gamma, and update frequency, were shown to substantially affect the DQN model’s learning ability. These findings suggest that exploration-driven GA strategies outperform GA approaches with limited exploration, underscoring the critical role of selection and crossover methods in enhancing DQN model efficiency and performance. Moreover, a mini case study on the CartPole environment revealed that even a 5% sensor dropout impaired the performance of a GA-optimised RL agent, while a 20% dropout almost entirely halted improvements.https://www.mdpi.com/2076-3417/15/4/2067genetic algorithmreinforcement learninghyperparameteroptimisation
spellingShingle Bartłomiej Brzęk
Barbara Probierz
Jan Kozak
Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
Applied Sciences
genetic algorithm
reinforcement learning
hyperparameter
optimisation
title Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
title_full Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
title_fullStr Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
title_full_unstemmed Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
title_short Exploration-Driven Genetic Algorithms for Hyperparameter Optimisation in Deep Reinforcement Learning
title_sort exploration driven genetic algorithms for hyperparameter optimisation in deep reinforcement learning
topic genetic algorithm
reinforcement learning
hyperparameter
optimisation
url https://www.mdpi.com/2076-3417/15/4/2067
work_keys_str_mv AT bartłomiejbrzek explorationdrivengeneticalgorithmsforhyperparameteroptimisationindeepreinforcementlearning
AT barbaraprobierz explorationdrivengeneticalgorithmsforhyperparameteroptimisationindeepreinforcementlearning
AT jankozak explorationdrivengeneticalgorithmsforhyperparameteroptimisationindeepreinforcementlearning