Reference : Metaheuristics-based Exploration Strategies for Multi-Objective Reinforcement Learning

	Document type :	Scientific congresses, symposiums and conference proceedings : Paper published in a book
	Discipline(s) :	Engineering, computing & technology : Computer science
	To cite this reference:	http://hdl.handle.net/10993/50373

Title :	Metaheuristics-based Exploration Strategies for Multi-Objective Reinforcement Learning
Language :	English
Author, co-author :	Felten, Florian [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > PCOG]
	Danoy, Grégoire [University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Computer Science (DCS) > ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > PCOG]
	Talbi, El-Ghazali [University of Lille, CNRS/CRIStAL, Inria Lille, France]
	Bouvry, Pascal [University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Computer Science (DCS) > ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > PCOG]
Publication date :	2022
Main document title :	Proceedings of the 14th International Conference on Agents and Artificial Intelligence
Publisher :	SCITEPRESS - Science and Technology Publications
Pages :	662--673
Peer reviewed :	Yes
On invitation :	No
Audience :	International
ISBN :	978-989-758-547-0
City :	Online Streaming
Event name :	14th International Conference on Agents and Artificial Intelligence
Event date :	from 3-02-2022 to 5-02-2022
Keywords :	[en] Reinforcement Learning ; Multi-objective ; Metaheuristics ; Pareto Sets
Abstract :	[en] The fields of Reinforcement Learning (RL) and Optimization aim at finding an optimal solution to a problem, characterized by an objective function. The exploration-exploitation dilemma (EED) is a well known subject in those fields. Indeed, a consequent amount of literature has already been proposed on the subject and shown it is a non-negligible topic to consider to achieve good performances. Yet, many problems in real life involve the optimization of multiple objectives. Multi-Policy Multi-Objective Reinforcement Learning (MPMORL) offers a way to learn various optimised behaviours for the agent in such problems. This work introduces a modular framework for the learning phase of such algorithms, allowing to ease the study of the EED in Inner- Loop MPMORL algorithms. We present three new exploration strategies inspired from the metaheuristics domain. To assess the performance of our methods on various environments, we use a classical benchmark - the Deep Sea Treasure (DST) - as well as propose a harder version of it. Our experiments show all of the proposed strategies outperform the current state-of-the-art ε-greedy based methods on the studied benchmarks.
Research centres :	Interdisciplinary Centre for Security, Reliability and Trust (SnT) > Parallel Computing & Optimization Group (PCOG)
Funders :	Fonds National de la Recherche - FnR
Target :	Researchers ; Professionals ; Students ; General public
Permalink :	http://hdl.handle.net/10993/50373
DOI :	10.5220/0010989100003116
Other URL :	https://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0010989100003116
FnR project :	FnR ; FNR14762457 > Gregoire Danoy > ADARS > Automating The Design Of Autonomous Robot Swarms > 01/05/2021 > 30/04/2024 > 2020

	File(s) associated to this reference
	Fulltext file(s): File Commentary Version Size Access Limited access 109891.pdf Publisher postprint 930.73 kB Request a copy

All documents in ORBi^lu are protected by a user license.

 

University of Luxembourg Library |  Feedback |  Legal notices Site Map