AAAI Conference 2004 System Paper
Multi-Agent System Development: Design, Runtime, and Analysis
- K. S. Barber
- K. Fullam
- N. Gujral
- D. N. Lam
- J. Park
Dynamic and unexpected events are defining characteristics of numerous application domains. These environments often require decision-makers to solve many problems with insufficient time and resources. To effectively assess options, decision-makers require situation analysis and decision-support tools that model the dynamism of these environments to make rapid, robust decisions. Autonomous agents and multi-agent systems satisfy these requirements for decision-making in dynamic environments. Developing a multi-agent system (MAS) is a challenging task, considering sophisticated agent interactions and uncertain environmental dynamics and domain requirements. This demonstration addresses the comprehensive development process for multi-agent systems; illustrating tools for the initial design of the agent system, the capabilities encoded in the individual agents, and analysis tools that enhance developer comprehension of system behavior.