Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning

Satoshi Kurihara, Toshiharu Sugawara, Rikio Onai

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper proposes and evaluates MarLee, a multi-agent reinforcement learning system that integrates both exploitation- and exploration-oriented learning. Compared with conventional reinforcement learnings, MarLee is more robust in the face of a dynamically changing environment and is able to perform exploration-oriented learning efficiently even in a large-scale environment. Thus, MarLee is well suited for autonomous systems, for example, software agents and mobile robots, that operate in dynamic, large-scale environments, like the real-world and the Internet. Spreading activation, based on the behavior-based approach, is used to explore the environment, so by manipulating the parameters of the spreading activation, it is easy to tune the learning characteristics. The fundamental effectiveness of MarLee was demonstrated by simulation.

Original languageEnglish
Title of host publicationMulti-Agent Systems
Subtitle of host publicationTheories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers
PublisherSpringer Verlag
Pages45-57
Number of pages13
ISBN (Print)3540654771, 9783540654773
Publication statusPublished - 1998 Jan 1
Externally publishedYes
Event4th Australian Workshop on Distributed Artificial Intelligence, DAK 1998 - Brisbane, Australia
Duration: 1998 Jul 131998 Jul 13

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1544
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other4th Australian Workshop on Distributed Artificial Intelligence, DAK 1998
CountryAustralia
CityBrisbane
Period98/7/1398/7/13

Fingerprint

Multiagent Learning
Reinforcement learning
Learning Systems
Reinforcement Learning
Exploitation
Learning systems
Chemical activation
Software agents
Mobile robots
Activation
Internet
Software Agents
Autonomous Systems
Mobile Robot
Integrate
Learning
Evaluate
Simulation

Keywords

  • Dynamic environment
  • Exploitation-oriented
  • Exploration-oriented
  • Multi-agent reinforcement learning
  • Spreading activation

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Kurihara, S., Sugawara, T., & Onai, R. (1998). Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning. In Multi-Agent Systems: Theories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers (pp. 45-57). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1544). Springer Verlag.

Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning. / Kurihara, Satoshi; Sugawara, Toshiharu; Onai, Rikio.

Multi-Agent Systems: Theories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers. Springer Verlag, 1998. p. 45-57 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1544).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kurihara, S, Sugawara, T & Onai, R 1998, Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning. in Multi-Agent Systems: Theories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1544, Springer Verlag, pp. 45-57, 4th Australian Workshop on Distributed Artificial Intelligence, DAK 1998, Brisbane, Australia, 98/7/13.
Kurihara S, Sugawara T, Onai R. Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning. In Multi-Agent Systems: Theories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers. Springer Verlag. 1998. p. 45-57. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Kurihara, Satoshi ; Sugawara, Toshiharu ; Onai, Rikio. / Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning. Multi-Agent Systems: Theories, Languages, and Applications - 4th Australian Workshop on Distributed Artificial Intelligence, 1998, Selected Papers. Springer Verlag, 1998. pp. 45-57 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{fbf5edc458cf45b0a7480283c0719742,
title = "Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning",
abstract = "This paper proposes and evaluates MarLee, a multi-agent reinforcement learning system that integrates both exploitation- and exploration-oriented learning. Compared with conventional reinforcement learnings, MarLee is more robust in the face of a dynamically changing environment and is able to perform exploration-oriented learning efficiently even in a large-scale environment. Thus, MarLee is well suited for autonomous systems, for example, software agents and mobile robots, that operate in dynamic, large-scale environments, like the real-world and the Internet. Spreading activation, based on the behavior-based approach, is used to explore the environment, so by manipulating the parameters of the spreading activation, it is easy to tune the learning characteristics. The fundamental effectiveness of MarLee was demonstrated by simulation.",
keywords = "Dynamic environment, Exploitation-oriented, Exploration-oriented, Multi-agent reinforcement learning, Spreading activation",
author = "Satoshi Kurihara and Toshiharu Sugawara and Rikio Onai",
year = "1998",
month = "1",
day = "1",
language = "English",
isbn = "3540654771",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "45--57",
booktitle = "Multi-Agent Systems",
address = "Germany",

}

TY - GEN

T1 - Multi-agent reinforcement learning system integrating exploitation- and exploration-oriented learning

AU - Kurihara, Satoshi

AU - Sugawara, Toshiharu

AU - Onai, Rikio

PY - 1998/1/1

Y1 - 1998/1/1

N2 - This paper proposes and evaluates MarLee, a multi-agent reinforcement learning system that integrates both exploitation- and exploration-oriented learning. Compared with conventional reinforcement learnings, MarLee is more robust in the face of a dynamically changing environment and is able to perform exploration-oriented learning efficiently even in a large-scale environment. Thus, MarLee is well suited for autonomous systems, for example, software agents and mobile robots, that operate in dynamic, large-scale environments, like the real-world and the Internet. Spreading activation, based on the behavior-based approach, is used to explore the environment, so by manipulating the parameters of the spreading activation, it is easy to tune the learning characteristics. The fundamental effectiveness of MarLee was demonstrated by simulation.

AB - This paper proposes and evaluates MarLee, a multi-agent reinforcement learning system that integrates both exploitation- and exploration-oriented learning. Compared with conventional reinforcement learnings, MarLee is more robust in the face of a dynamically changing environment and is able to perform exploration-oriented learning efficiently even in a large-scale environment. Thus, MarLee is well suited for autonomous systems, for example, software agents and mobile robots, that operate in dynamic, large-scale environments, like the real-world and the Internet. Spreading activation, based on the behavior-based approach, is used to explore the environment, so by manipulating the parameters of the spreading activation, it is easy to tune the learning characteristics. The fundamental effectiveness of MarLee was demonstrated by simulation.

KW - Dynamic environment

KW - Exploitation-oriented

KW - Exploration-oriented

KW - Multi-agent reinforcement learning

KW - Spreading activation

UR - http://www.scopus.com/inward/record.url?scp=84961359588&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961359588&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84961359588

SN - 3540654771

SN - 9783540654773

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 45

EP - 57

BT - Multi-Agent Systems

PB - Springer Verlag

ER -