Multiagent Coordination, Planning, and Control
An autonomous agent should coordinate its actions with the actions of
other agents to avoid unnecessary conflict and to exploit
cooperative opportunities. This requires that agents know enough about
what others are doing and plan to do to anticipate how
control decisions made now could impact the prospects of future
successes. In this tutorial, we will study strategies by which
agents can cooperatively solve problems of coordination and control in a
practical manner. Topics will include problem
representation and decomposition, task and resource allocation, conflict
detection and resolution, multi-agent plan improvement, and
probabilistic (Markov Decision Process) methods for coordinated
distributed control by resource-limited agents.
Biography
Edmund H. Durfee is a Professor at the University of Michigan, where he
has been on the faculty for over 15 years. He has conducted
research and published extensively in the areas of coordination in
multi-agent systems and real-time intelligent control. He is an
associate editor of the Journal of Autonomous Agents and Multi-agent
Systems, and is a Fellow of IEEE and of AAAI.
