Digital Library

cab1

 
Title:      THE PRESENTER ROLE MONITORING CAPABILITY IN A MULTI-AGENT ENVIRONMENT
Author(s):      Paulo T. Silva , Luís Moniz , Helder Coelho
ISBN:      972-99353-6-X
Editors:      Nuno Guimarães and Pedro Isaías
Year:      2005
Edition:      1
Keywords:      Multi-agent systems, ontologies, knowledge-level presentation.
Type:      Full Paper
First Page:      385
Last Page:      392
Language:      English
Cover:      cover          
Full Contents:      click to dowload Download
Paper Abstract:      This paper focuses on agent-to-human presentation of the agent knowledge-level in a multi agent environment. The presentation skill is regarded as an agent's role - the presenter role, in charge of uncovering what is occurring. In our approach the design of this role is driven by two main concerns: i) the inter-agent knowledge sharing, and ii) the human user monitoring of the knowledge being shared by the agents. In this paper we address these concerns from the agent’s ontology commitment perspective. We propose a presenter role model (and its subsequent architecture) centered on the domain and role ontologies. The implementation of the presenter role faces two challenges: one is to cope with diverse knowledge-level representations; the other is to present behavioral knowledge to the human user. In this paper we describe our approach to the implementation challenges and to the materialization of the presenter role model in the context of a multi-agent game-like experiment. In the experiment, the inter-agent communication uses a standard semantic language (FIPA SL) and the agent-to-user presentation adopts the Java Swing platform independent user-interface components. We adopted JADE (FIPA compliant) framework as the basis of the development process. Ontology specifications were built with Protégé ontology editor. We are currently exploring the presenter role monitoring capability in the RoboCupRescue multi-agent environment for simulations in the domain of emergency scenarios of search and rescue behavior within large-scale disasters.
   

Social Media Links

Search

Login