Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Autoren

  • Joshua Riley
  • Radu Calinescu
  • Colin Paterson
  • Daniel Kudenko
  • Alec Banks

Organisationseinheiten

Externe Organisationen

  • University of York
  • Defence Science and Technology Laboratory
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des SammelwerksIntelligent Decision Technologies
UntertitelProceedings of the 14th KES-IDT 2022 Conference
Herausgeber/-innenIreneusz Czarnowski, Robert J. Howlett, Robert J. Howlett, Lakhmi C. Jain
ErscheinungsortSingapore
Seiten87-97
Seitenumfang11
ISBN (elektronisch)978-981-19-3444-5
PublikationsstatusVeröffentlicht - 27 Juli 2022
Veranstaltung14th International KES Conference on Intelligent Decision Technologies, KES-IDT 2022 - Virtual, Online
Dauer: 20 Juni 202222 Juni 2022

Publikationsreihe

NameSmart Innovation, Systems and Technologies
Band309
ISSN (Print)2190-3018
ISSN (elektronisch)2190-3026

Abstract

Multi-agent reinforcement learning facilitates agents learning to solve complex decision-making problems requiring collaboration. However, reinforcement learning methods are underpinned by stochastic mechanisms, making them unsuitable for safety-critical domains. To solve this issue, approaches such as assured multi-agent reinforcement learning, which utilises quantitative verification to produce formal guarantees of safety requirements during the agents learning process, have been developed. However, this approach relies on accurate knowledge about the environment to be effectively used which can be detrimental if this knowledge is inaccurate. Therefore, we developed an extension to assured multi-agent reinforcement learning called agent interaction driven adaptability, an automated process to securing reliable safety constraints, allowing inaccurate and missing knowledge to be used without detriment. Our preliminary results showcase the ability of agent interaction driven adaptability to allow safe multi-agent reinforcement learning to be utilised in safety-critical scenarios.

ASJC Scopus Sachgebiete

Zitieren

Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability. / Riley, Joshua; Calinescu, Radu; Paterson, Colin et al.
Intelligent Decision Technologies : Proceedings of the 14th KES-IDT 2022 Conference. Hrsg. / Ireneusz Czarnowski; Robert J. Howlett; Robert J. Howlett; Lakhmi C. Jain. Singapore, 2022. S. 87-97 (Smart Innovation, Systems and Technologies; Band 309).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Riley, J, Calinescu, R, Paterson, C, Kudenko, D & Banks, A 2022, Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability. in I Czarnowski, RJ Howlett, RJ Howlett & LC Jain (Hrsg.), Intelligent Decision Technologies : Proceedings of the 14th KES-IDT 2022 Conference. Smart Innovation, Systems and Technologies, Bd. 309, Singapore, S. 87-97, 14th International KES Conference on Intelligent Decision Technologies, KES-IDT 2022, Virtual, Online, 20 Juni 2022. https://doi.org/10.1007/978-981-19-3444-5_8
Riley, J., Calinescu, R., Paterson, C., Kudenko, D., & Banks, A. (2022). Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability. In I. Czarnowski, R. J. Howlett, R. J. Howlett, & L. C. Jain (Hrsg.), Intelligent Decision Technologies : Proceedings of the 14th KES-IDT 2022 Conference (S. 87-97). (Smart Innovation, Systems and Technologies; Band 309).. https://doi.org/10.1007/978-981-19-3444-5_8
Riley J, Calinescu R, Paterson C, Kudenko D, Banks A. Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability. in Czarnowski I, Howlett RJ, Howlett RJ, Jain LC, Hrsg., Intelligent Decision Technologies : Proceedings of the 14th KES-IDT 2022 Conference. Singapore. 2022. S. 87-97. (Smart Innovation, Systems and Technologies). doi: 10.1007/978-981-19-3444-5_8
Riley, Joshua ; Calinescu, Radu ; Paterson, Colin et al. / Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability. Intelligent Decision Technologies : Proceedings of the 14th KES-IDT 2022 Conference. Hrsg. / Ireneusz Czarnowski ; Robert J. Howlett ; Robert J. Howlett ; Lakhmi C. Jain. Singapore, 2022. S. 87-97 (Smart Innovation, Systems and Technologies).
Download
@inproceedings{fcf6b0bcc3734ff39ff36fc4364316f9,
title = "Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability",
abstract = "Multi-agent reinforcement learning facilitates agents learning to solve complex decision-making problems requiring collaboration. However, reinforcement learning methods are underpinned by stochastic mechanisms, making them unsuitable for safety-critical domains. To solve this issue, approaches such as assured multi-agent reinforcement learning, which utilises quantitative verification to produce formal guarantees of safety requirements during the agents learning process, have been developed. However, this approach relies on accurate knowledge about the environment to be effectively used which can be detrimental if this knowledge is inaccurate. Therefore, we developed an extension to assured multi-agent reinforcement learning called agent interaction driven adaptability, an automated process to securing reliable safety constraints, allowing inaccurate and missing knowledge to be used without detriment. Our preliminary results showcase the ability of agent interaction driven adaptability to allow safe multi-agent reinforcement learning to be utilised in safety-critical scenarios.",
keywords = "Assurance, Deep reinforcement learning, Multi-agent reinforcement learning, Quantitative verification, Safety",
author = "Joshua Riley and Radu Calinescu and Colin Paterson and Daniel Kudenko and Alec Banks",
year = "2022",
month = jul,
day = "27",
doi = "10.1007/978-981-19-3444-5_8",
language = "English",
isbn = "9789811934438",
series = "Smart Innovation, Systems and Technologies",
pages = "87--97",
editor = "Ireneusz Czarnowski and Howlett, {Robert J.} and Howlett, {Robert J.} and Jain, {Lakhmi C.}",
booktitle = "Intelligent Decision Technologies",
note = "14th International KES Conference on Intelligent Decision Technologies, KES-IDT 2022 ; Conference date: 20-06-2022 Through 22-06-2022",

}

Download

TY - GEN

T1 - Assured Multi-agent Reinforcement Learning with Robust Agent-Interaction Adaptability

AU - Riley, Joshua

AU - Calinescu, Radu

AU - Paterson, Colin

AU - Kudenko, Daniel

AU - Banks, Alec

PY - 2022/7/27

Y1 - 2022/7/27

N2 - Multi-agent reinforcement learning facilitates agents learning to solve complex decision-making problems requiring collaboration. However, reinforcement learning methods are underpinned by stochastic mechanisms, making them unsuitable for safety-critical domains. To solve this issue, approaches such as assured multi-agent reinforcement learning, which utilises quantitative verification to produce formal guarantees of safety requirements during the agents learning process, have been developed. However, this approach relies on accurate knowledge about the environment to be effectively used which can be detrimental if this knowledge is inaccurate. Therefore, we developed an extension to assured multi-agent reinforcement learning called agent interaction driven adaptability, an automated process to securing reliable safety constraints, allowing inaccurate and missing knowledge to be used without detriment. Our preliminary results showcase the ability of agent interaction driven adaptability to allow safe multi-agent reinforcement learning to be utilised in safety-critical scenarios.

AB - Multi-agent reinforcement learning facilitates agents learning to solve complex decision-making problems requiring collaboration. However, reinforcement learning methods are underpinned by stochastic mechanisms, making them unsuitable for safety-critical domains. To solve this issue, approaches such as assured multi-agent reinforcement learning, which utilises quantitative verification to produce formal guarantees of safety requirements during the agents learning process, have been developed. However, this approach relies on accurate knowledge about the environment to be effectively used which can be detrimental if this knowledge is inaccurate. Therefore, we developed an extension to assured multi-agent reinforcement learning called agent interaction driven adaptability, an automated process to securing reliable safety constraints, allowing inaccurate and missing knowledge to be used without detriment. Our preliminary results showcase the ability of agent interaction driven adaptability to allow safe multi-agent reinforcement learning to be utilised in safety-critical scenarios.

KW - Assurance

KW - Deep reinforcement learning

KW - Multi-agent reinforcement learning

KW - Quantitative verification

KW - Safety

UR - http://www.scopus.com/inward/record.url?scp=85135905227&partnerID=8YFLogxK

U2 - 10.1007/978-981-19-3444-5_8

DO - 10.1007/978-981-19-3444-5_8

M3 - Conference contribution

AN - SCOPUS:85135905227

SN - 9789811934438

T3 - Smart Innovation, Systems and Technologies

SP - 87

EP - 97

BT - Intelligent Decision Technologies

A2 - Czarnowski, Ireneusz

A2 - Howlett, Robert J.

A2 - Howlett, Robert J.

A2 - Jain, Lakhmi C.

CY - Singapore

T2 - 14th International KES Conference on Intelligent Decision Technologies, KES-IDT 2022

Y2 - 20 June 2022 through 22 June 2022

ER -