Modeling a distributed multi-agent system to agent distributing based on behavior patterns

Juan Carlos Vidal, José Miguel Piquer

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

Abstract

Inside a distributed Multi-Agent Systems DMAS happen several events during agent's life cycle that changing dynamically the state of each agent and all DMAS. This dynamic nature is based on interchanged information between agents that support the cooperation to obtain a commun objective. Moreover in the optimization of distributed Multi-Agent Systems raises the issue of agent distributing according to the availability of resources. This paper presents a agent distributing strategy for DMAS based on a predictive system based on agent behavior patterns modeling. Our strategy allows the prediction of the impact of the creation of new agents on the local nodes and, hence making of optimal decisions with the purpose of improving future interactions. We present several analysis based on the execution of agents groups and the results show that a DMAS is improved.

Original languageEnglish
Title of host publicationProceedings of the Seventh IASTED International Conference on Internet and Multimedia Systems and Applications
EditorsM.H. Hamza
Pages717-722
Number of pages6
StatePublished - 2003
Externally publishedYes
EventProceedings of the Seventh IASTED International Conference on Internet and Multimedia Systems and Applications - Honolulu, HI, United States
Duration: 13 Aug 200315 Aug 2003

Publication series

NameProceedings of the IASTED International Conference on Internet and Multimedia Systems and Applications
Volume7

Conference

ConferenceProceedings of the Seventh IASTED International Conference on Internet and Multimedia Systems and Applications
Country/TerritoryUnited States
CityHonolulu, HI
Period13/08/0315/08/03

Keywords

  • Agent dynamic distributing
  • Multi-Agent System Modeling
  • Predictive System
  • Volatile agent

Fingerprint

Dive into the research topics of 'Modeling a distributed multi-agent system to agent distributing based on behavior patterns'. Together they form a unique fingerprint.

Cite this