Skip to main navigation Skip to search Skip to main content

Utility-Based Sequential Decision-Making In Evidential Cooperative Multi-Agent Systems

  • SUNY Buffalo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

This paper presents a new approach to building utili&-based models of decision-makingin time-constrainedsituations with limited resources. A particular hierarchical homogenousmulti-agentarchitecture has been considered. Theproposed system combines agents’ belieji within the frame work of evidence theory and afler each observation maps the current set of cumulativepignistic probabilities into one of two actions: “defer decision” or “decide hypothesis i ”. The system maximizes the expected utility of delayed decisions minus cost. Theprocess of system adaptation to the environment is guided by reinforcement learning. The utilities-from-experts problem is simplified by learning utilities directly from feedbackon the qualiyof the decisions. The results of a case study are presented.

Original languageEnglish
Title of host publicationProceedings of the 6th International Conference on Information Fusion, FUSION 2003
PublisherIEEE Computer Society
Pages823-830
Number of pages8
ISBN (Print)0972184449, 9780972184441
DOIs
StatePublished - 2003
Event6th International Conference on Information Fusion, FUSION 2003 - Cairns, QLD, Australia
Duration: Jul 8 2003Jul 11 2003

Publication series

NameProceedings of the 6th International Conference on Information Fusion, FUSION 2003
Volume2

Conference

Conference6th International Conference on Information Fusion, FUSION 2003
Country/TerritoryAustralia
CityCairns, QLD
Period07/8/0307/11/03

Keywords

  • Decision utility
  • Distributed systems
  • Evidence theory
  • Multi-agent systems
  • Reinforcement learning
  • Sequential decision-making

Fingerprint

Dive into the research topics of 'Utility-Based Sequential Decision-Making In Evidential Cooperative Multi-Agent Systems'. Together they form a unique fingerprint.

Cite this