Department of Statistical Science
Duke University

presents:

Michael Littman
Computer Science, Duke University

"Decision-Making Under Uncertainty Via Partially Observable Markov Decision Processes"

Abstract:

Sequential decision making is a fundamental task faced by any intelligent agent in an extended interaction with its environment. I have examined the problem of making optimal decisions in partially observable Markov decision processes (POMDPs), a type of sequential problem in which the state of the system and its future evolution are modeled stochastically. I will describe the POMDP model and an algorithm I developed for solving POMDPs exactly; I will show how it compares favorably to existing algorithms for this problem. I will also present some preliminary results on the use of a reinforcement-learning approach to solve larger problem instances approximately; this technique has been used successfully to find good policies for a robot navigating in an office environment.

Friday, October 25, 1996

4:00 pm - 5:00 pm

116 Old Chem Building

Any questions concerning the seminar may be addressed to Cheryl McGhee @[919] 684-8029, e-mail cheryl@stat.duke.edu