Journal article

The Optimal Observability of Partially Observable Markov Decision Processes: Discrete State Space

Mohammad Rezaeian, Ba-Ngu Vo, Jamie Scott Evans

IEEE TRANSACTIONS ON AUTOMATIC CONTROL | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC | Published : 2010

Abstract

We consider autonomous partially observable Markov decision processes where the control action influences the observation process only. Considering entropy as the cost incurred by the Markov information state process, the optimal observability problem is posed as a Markov decision scheduling problem that minimizes the infinite horizon cost. This scheduling problem is shown to be equivalent to minimization of an entropy measure, called estimation entropy which is related to the invariant measure of the information state. © 2006 IEEE.

University of Melbourne Researchers

Grants

Awarded by Australian Research Council


Funding Acknowledgements

Manuscript received January 13, 2009; revised August 27, 2009, March 08, 2010, and June 10, 2010; accepted July 06, 2010. Date of publication September 09, 2010; date of current version December 02, 2010. This work was supported by the Australian Research Council, Discovery Grant DP0878158. Recommended by Associate Editor C. Szepesvari.