Publications‎ > ‎

Design, Runtime, and Analysis of Multi-Agent Systems


Abstract:


Developing a multi-agent system (MAS) is a challenging task, considering sophisticated agent interactions, uncertain environmental conditions, and dynamic domain requirements. This set of demonstrations offers tools for the design and analysis of agent systems and implementations. The design tools help to quickly select agent functionality, assign functionality to components, and then evaluate agent technologies for potential inclusion in the agent design. Search and evaluation is based on the ability of candidate technologies to deliver required functionality and satisfy deployment constraints (e.g. interoperability with database choice, operation on selected OS, etc). The analysis tool, Tracer, is used to validate the implemented behaviors against the designed behaviors and can be used for software comprehension. Sample technologies illustrated in this work include novel work in belief revision, information source selection, and action selection. These technologies are encoded in each agent to form a reliable information network and to support agents' action-selection in an Unmanned Aerial Vehicle (UAV) target tracking simulation. 


Park, J., J. Ahn, D. DeAngelis, K. Fullam, N. Gujral, D. Han, D. Lam, and K. S. Barber. Design, Runtime, and Analysis of Multi-Agent Systems. In the Proceedings of The Fourth International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS-2005) Demonstration Track; Utrecht, The Netherlands; July 25-29, 2005; pp. 157-158.

Ċ
Dave DeAngelis,
Jun 4, 2009, 1:32 PM
Comments