An approach to solving deadlock in multi-issue negotiation

Author

Qing, Guo ; Chun, Chen

Author_Institution

Dept. of Comput. Sci., Zhejiang Univ., Hangzhou, China

fYear

2003

fDate

27-29 Oct. 2003

Firstpage

48

Lastpage

52

Abstract

In multi-issue negotiation among agents, how to resolve negotiation failure induced by deadlock involved with one certain issue is an inviting task. We propose an equivalent replacement mechanism for reservation values of multiple issues in bi-lateral negotiation, which dynamically adjusts reservation values of every issue to solve negotiation deadlock on the premise of guaranteeing the integrated utility of the participants, and give a tactic of equivalent replacement for reservation valued based on reinforcement-learning algorithm, which maximizes the integrated utility of participants and improves the negotiation efficiency.

Keywords

learning (artificial intelligence); multi-agent systems; bi-lateral negotiation; multiagent system; multiissue agent negotiation; negotiation deadlock; negotiation failure; reinforcement learning; reinforcement-learning algorithm; replacement mechanism; reservation value adjustment; reservation values; Bayesian methods; Computer science; Costs; Game theory; Genetic algorithms; Iterative algorithms; Machine learning algorithms; Multiagent systems; Protocols; System recovery;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Reuse and Integration, 2003. IRI 2003. IEEE International Conference on

Print_ISBN

0-7803-8242-0

Type

conf

DOI

10.1109/IRI.2003.1251394

Filename

1251394