Strategy acquisition on multi-issue negotiation without estimating opponent's preference

Shohei Yoshikawa, Yoshiaki Yasumura, Kuniaki Uehara

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

1 Citation (Scopus)

Abstract

In multi-issue negotiation, an opponent's preference is rarely open. Under this environment, it is difficult to acquire a negotiation result that realizes win-win negotiation. In this paper, we present a novel method for realizing win-win negotiation although an opponent's preference is not open. In this method, an agent learns how to make a concession to an opponent. To learn the concession strategy, we adopt reinforcement learning. In reinforcement learning, the agent recognizes a negotiation state to each issue in negotiation. According to the state, the agent makes a proposal to increase own profit. A reward of the learning is a profit of an agreement and punishment of negotiation breakdown. Experimental results showed that agents could acquire a negotiation strategy that avoids negotiation breakdown and increases profits of both sides. Finally, the agents can acquire the action policy that strikes a balance between cooperation and competition.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages371-380
Number of pages10
Volume4953 LNAI
DOIs
Publication statusPublished - 2008
Externally publishedYes
Event2nd KES International Symposium on Agent and Multi-Agent Systems: Technologies and Applications, KES-AMSTA 2008 - Incheon
Duration: 2008 Mar 262008 Mar 28

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4953 LNAI
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other2nd KES International Symposium on Agent and Multi-Agent Systems: Technologies and Applications, KES-AMSTA 2008
CityIncheon
Period08/3/2608/3/28

Fingerprint

Profitability
Reinforcement learning

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Yoshikawa, S., Yasumura, Y., & Uehara, K. (2008). Strategy acquisition on multi-issue negotiation without estimating opponent's preference. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4953 LNAI, pp. 371-380). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4953 LNAI). https://doi.org/10.1007/978-3-540-78582-8_38

Strategy acquisition on multi-issue negotiation without estimating opponent's preference. / Yoshikawa, Shohei; Yasumura, Yoshiaki; Uehara, Kuniaki.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4953 LNAI 2008. p. 371-380 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4953 LNAI).

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

Yoshikawa, S, Yasumura, Y & Uehara, K 2008, Strategy acquisition on multi-issue negotiation without estimating opponent's preference. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 4953 LNAI, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4953 LNAI, pp. 371-380, 2nd KES International Symposium on Agent and Multi-Agent Systems: Technologies and Applications, KES-AMSTA 2008, Incheon, 08/3/26. https://doi.org/10.1007/978-3-540-78582-8_38
Yoshikawa S, Yasumura Y, Uehara K. Strategy acquisition on multi-issue negotiation without estimating opponent's preference. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4953 LNAI. 2008. p. 371-380. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-540-78582-8_38
Yoshikawa, Shohei ; Yasumura, Yoshiaki ; Uehara, Kuniaki. / Strategy acquisition on multi-issue negotiation without estimating opponent's preference. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4953 LNAI 2008. pp. 371-380 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{092621a721954cf6a16c798d4556afd4,
title = "Strategy acquisition on multi-issue negotiation without estimating opponent's preference",
abstract = "In multi-issue negotiation, an opponent's preference is rarely open. Under this environment, it is difficult to acquire a negotiation result that realizes win-win negotiation. In this paper, we present a novel method for realizing win-win negotiation although an opponent's preference is not open. In this method, an agent learns how to make a concession to an opponent. To learn the concession strategy, we adopt reinforcement learning. In reinforcement learning, the agent recognizes a negotiation state to each issue in negotiation. According to the state, the agent makes a proposal to increase own profit. A reward of the learning is a profit of an agreement and punishment of negotiation breakdown. Experimental results showed that agents could acquire a negotiation strategy that avoids negotiation breakdown and increases profits of both sides. Finally, the agents can acquire the action policy that strikes a balance between cooperation and competition.",
author = "Shohei Yoshikawa and Yoshiaki Yasumura and Kuniaki Uehara",
year = "2008",
doi = "10.1007/978-3-540-78582-8_38",
language = "English",
isbn = "3540785817",
volume = "4953 LNAI",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "371--380",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Strategy acquisition on multi-issue negotiation without estimating opponent's preference

AU - Yoshikawa, Shohei

AU - Yasumura, Yoshiaki

AU - Uehara, Kuniaki

PY - 2008

Y1 - 2008

N2 - In multi-issue negotiation, an opponent's preference is rarely open. Under this environment, it is difficult to acquire a negotiation result that realizes win-win negotiation. In this paper, we present a novel method for realizing win-win negotiation although an opponent's preference is not open. In this method, an agent learns how to make a concession to an opponent. To learn the concession strategy, we adopt reinforcement learning. In reinforcement learning, the agent recognizes a negotiation state to each issue in negotiation. According to the state, the agent makes a proposal to increase own profit. A reward of the learning is a profit of an agreement and punishment of negotiation breakdown. Experimental results showed that agents could acquire a negotiation strategy that avoids negotiation breakdown and increases profits of both sides. Finally, the agents can acquire the action policy that strikes a balance between cooperation and competition.

AB - In multi-issue negotiation, an opponent's preference is rarely open. Under this environment, it is difficult to acquire a negotiation result that realizes win-win negotiation. In this paper, we present a novel method for realizing win-win negotiation although an opponent's preference is not open. In this method, an agent learns how to make a concession to an opponent. To learn the concession strategy, we adopt reinforcement learning. In reinforcement learning, the agent recognizes a negotiation state to each issue in negotiation. According to the state, the agent makes a proposal to increase own profit. A reward of the learning is a profit of an agreement and punishment of negotiation breakdown. Experimental results showed that agents could acquire a negotiation strategy that avoids negotiation breakdown and increases profits of both sides. Finally, the agents can acquire the action policy that strikes a balance between cooperation and competition.

UR - http://www.scopus.com/inward/record.url?scp=47249116495&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47249116495&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-78582-8_38

DO - 10.1007/978-3-540-78582-8_38

M3 - Conference contribution

AN - SCOPUS:47249116495

SN - 3540785817

SN - 9783540785811

VL - 4953 LNAI

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

SP - 371

EP - 380

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

ER -