See detailed view

Peer-reviewed conferences/proceedings

Symposium on Applied Computing - SAC'2001, Las Vegas, USA : (2001)

A New Decomposition Technique for Solving Markov Decision Processes

Pierre Laroche1, Yann Boniface2, René Schott3, 4

In this paper, we present a new tool for automatically solving Markov Decision Processes. Using a predefined partition o fthe MDP, a directed graph is built to decompose the global MDP into small local MDPs which are independently solved. An approximate solution for the global MDP is obtained using local solutions. Our approach has been tested on a mobile robotics application. It allows near-optimal solutions to be obtained in significantly reduced time. We also present preliminary results concerning a parallel implementation.

1:	LITA (EA3097) - Laboratoire d'Informatique Théorique et Appliquée
2:	INRIA Lorraine - LORIA - CORTEX
3:	LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
4:	IECN - Institut Elie Cartan Nancy

Keywords : planning under uncertainty – markov decision process – decomposition – parallelism || plannification sous incertitude – mdp – parallélisme

inria-00101093, version 1

http://hal.inria.fr/inria-00101093

oai:hal.inria.fr:inria-00101093

From: Publications Loria <>

Submitted on: Tuesday, 26 September 2006 14:56:27

Updated on: Monday, 11 June 2007 15:37:31