GMDPtoolbox proposes functions related to Graph-based Markov Decision Processes (GMDP). The framework allows to represent and approximately solve Markov Decision Processes (MDP) problems with an underlying spatial structure allowing a factored representation. A factored representation can be a solution for problems too large to be represented and solved by classical MDP tools. For smaller size problems, we suggest to use the MDP toolbox.
The toolbox is under BSD license and currently available for MATLAB.
See the Install page for details, clicking on the download icon on top.
See the Documentation page, clicking on the document icon on top.
MJ. Cros, JN. Aubertot, N. Peyrard and R. Sabbadin. GMDPtoolbox: A Matlab library for designing spatial management policies. Application to the long-term collective management of an airborne disease. PLoS ONE 12(10), 2017. [paper]
MJ. Cros, N. Peyrard and R. Sabbadin. GMDPtoolbox: a Matlab library for solving Graph-based Markov Decision Processes. Journées Francophones sur les Réseaux Bayésiens et les Modèles Graphiques Probabilistes (JFRB), Clermont-Ferrand France, 27-28 juin 2016. [paper][slides]>
N. Forsell, N. Peyrard, R. Sabbadin. A framework and a mean-field algorithm for the local control of spatial processes. International Journal of Approximate Reasoning, p 66-86, 2012. [paper]
The GMDP framework and the toolbox are developed by MJ. Cros, N. Peyrard and R. Sabbadin, MIAT Laboratory, INRA Toulouse, France.
The toolbox has received funding from the European union seventh Framework Programme (FP7/2007-2013) under the grant agreement n°265865.