Staff View: Re-exploration of ε-greedy in deep reinforcement learning

Re-exploration of ε-greedy in deep reinforcement learning

This paper presents re-exploration as a method for improving the existing method for balancing the exploration/exploitation problem integral to reinforcement learning. The proposed method uses a ε-greedy method called “decreasing epsilon,” which reiterate the method after a certain period of episode...

Full description

Saved in:

Bibliographic Details
Main Authors:	Muhamad Amin, Muhamad Ridzuan Radin, Othman, Mohd. Fauzi
Format:	Conference or Workshop Item
Published:	2021
Subjects:	L Education (General) TA Engineering (General). Civil engineering (General)
Online Access:	http://eprints.utm.my/id/eprint/96174/ http://dx.doi.org/10.1007/978-981-16-4803-8_27
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my.utm.96174
record_format	eprints
spelling	my.utm.961742022-07-04T07:57:36Z http://eprints.utm.my/id/eprint/96174/ Re-exploration of ε-greedy in deep reinforcement learning Muhamad Amin, Muhamad Ridzuan Radin Othman, Mohd. Fauzi L Education (General) TA Engineering (General). Civil engineering (General) This paper presents re-exploration as a method for improving the existing method for balancing the exploration/exploitation problem integral to reinforcement learning. The proposed method uses a ε-greedy method called “decreasing epsilon,” which reiterate the method after a certain period of episodes in the middle of the learning. The experiment was conducted using Turtlebot3 simulation under the Robot Operating System (ROS) environment. The evaluation involved comparing the existing method, which is pure exploitation (totally greedy), conventional ε-greedy method and proposed method, which is decreasing-epsilon with the re-exploration method. The preliminary results indicate that applying re-exploration method is easier to implement and yet able to improve the reward obtained with in shorter time (episode) compared to the conventional method. 2021 Conference or Workshop Item PeerReviewed Muhamad Amin, Muhamad Ridzuan Radin and Othman, Mohd. Fauzi (2021) Re-exploration of ε-greedy in deep reinforcement learning. In: 8th International Conference on Robot Intelligence Technology and Applications, RiTA 2020, 11 December 2020 - 13 December 2020, Virtual, Online. http://dx.doi.org/10.1007/978-981-16-4803-8_27
institution	Universiti Teknologi Malaysia
building	UTM Library
collection	Institutional Repository
continent	Asia
country	Malaysia
content_provider	Universiti Teknologi Malaysia
content_source	UTM Institutional Repository
url_provider	http://eprints.utm.my/
topic	L Education (General) TA Engineering (General). Civil engineering (General)
spellingShingle	L Education (General) TA Engineering (General). Civil engineering (General) Muhamad Amin, Muhamad Ridzuan Radin Othman, Mohd. Fauzi Re-exploration of ε-greedy in deep reinforcement learning
description	This paper presents re-exploration as a method for improving the existing method for balancing the exploration/exploitation problem integral to reinforcement learning. The proposed method uses a ε-greedy method called “decreasing epsilon,” which reiterate the method after a certain period of episodes in the middle of the learning. The experiment was conducted using Turtlebot3 simulation under the Robot Operating System (ROS) environment. The evaluation involved comparing the existing method, which is pure exploitation (totally greedy), conventional ε-greedy method and proposed method, which is decreasing-epsilon with the re-exploration method. The preliminary results indicate that applying re-exploration method is easier to implement and yet able to improve the reward obtained with in shorter time (episode) compared to the conventional method.
format	Conference or Workshop Item
author	Muhamad Amin, Muhamad Ridzuan Radin Othman, Mohd. Fauzi
author_facet	Muhamad Amin, Muhamad Ridzuan Radin Othman, Mohd. Fauzi
author_sort	Muhamad Amin, Muhamad Ridzuan Radin
title	Re-exploration of ε-greedy in deep reinforcement learning
title_short	Re-exploration of ε-greedy in deep reinforcement learning
title_full	Re-exploration of ε-greedy in deep reinforcement learning
title_fullStr	Re-exploration of ε-greedy in deep reinforcement learning
title_full_unstemmed	Re-exploration of ε-greedy in deep reinforcement learning
title_sort	re-exploration of ε-greedy in deep reinforcement learning
publishDate	2021
url	http://eprints.utm.my/id/eprint/96174/ http://dx.doi.org/10.1007/978-981-16-4803-8_27
_version_	1738510333221273600
score	13.160551

Re-exploration of ε-greedy in deep reinforcement learning

Similar Items