Seminars/Colloquia/Invited Talks

Seminars

Stephen Patek

Partially Observed Stochastic Shortest Path Problems with Approximate Solution by Neuro-Dynamic Programming

PLACE:	Clark 314
EVENT:	CIS Seminar Series
DATE:	November 30, 2004
TIME:	1:00 - 2:00

Abstract

In this talk, I analyze a class of Markov decision processes with imperfect state information that evolve on an infinite time horizon and have a total cost criterion. After making standard “stochastic shortest path” assumption, i.e. (1) the existence of a policy that guarantees termination with probability one and (2) the property that any policy that fails to guarantee termination has infinite expected cost from some initial state, I show that standard optimality equations characterize stationary optimal policies as long as termination is perfectly recognized. To conclude, I will present an illustrative example that involves the search for a partially observed target that moves randomly on a grid, with approximate solution based on neuro-dynamic programming

Brief biography

Stephen Patek is an Associate Professor in the Department of Systems and Information Engineering at the University of Virginia. His research interests include stochastic control and game theoretic models for communication and control of data and sensor networks. He attended graduate school at the Massachusetts Institute of Technology, in the Laboratory for Information and Decision Systems and Department of Electrical Engineering and Computer Science, finishing in 1997. He received his Bachelor’s degree in Electrical and Computer Engineering in 1991 from the University of Tennessee, Knoxville.