[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Reminder: Seminar today by Dana Nau (11:30AM BY210) on "Planning for Interactions among Autonomous Agents"



Planning for Interactions among Autonomous Agents

Dana S. Nau
Department of Computer Science
University of Maryland
College Park, MD

Date: Monday, March 31, 2008
Time: 11:30 AM - 12:45 PM
Place: Brickyard Building, Room 210

This talk will focus on ways to plan an autonomous agent's
interactions with other autonomous agents. Sometimes it is feasible to
model the other agents' possible actions as nondeterministic outcomes
of our agent's own actions. For this case, we can plan how to achieve
a desired goal using symbolic model checking, HTN planning, or a
combination of the two. Sometimes it may not be feasible to generate a
plan or policy that goes all the way to a goal, either because the
search space is too large or the goal is ambiguously defined. For such
cases, it can work well to interleave planning and execution if we
have a good predictive model of how the other agents are likely to
behave. The talk will present theoretical foundations and algorithms
for the above situations, and experimental results on the
Hunter-and-Prey domain, the Iterated Prisoner's Dilemma with Noise,
and other multi-agent planning domains.


Biography

Dana Nau is a Professor of both Computer Science and Systems Research
at the University of Maryland, and is co-director of the university's
Laboratory for Computational Cultural Dynamics.He has more than 300
refereed technical publications on automated planning, search
algorithms, game theory, and other topics.He has received an NSF PYI
award, several best-paper awards, and several prizes for the
performance of software systems.He is a Fellow of the Association for
the Advancement of Artificial Intelligence (AAAI).

==============