Fix V initialization in PolicyIterationModified

sawcordwell / pymdptoolbox

Markov Decision Process (MDP) Toolbox for Python

BSD 3-Clause "New" or "Revised" License

521 stars 251 forks source link

Fix V initialization in PolicyIterationModified #5

Closed yasserglez closed 9 years ago

yasserglez commented 9 years ago

V was initialized with the wrong dimensions for undiscounted MDPs in PolicyIterationModified, which caused an exception to be raised in _bellmanOperator.

coveralls commented 9 years ago

Coverage remained the same when pulling e997f549e35e747f3d9ef3ae7c8dcf2f9c690daa on yasserglez:fix-V-init-undiscounted into 3d20ccb4589abc203dc40b4e427a71e2a85337d0 on sawcordwell:master.

sawcordwell commented 9 years ago

Thanks, there really needs to be some unit tests for undiscounted MDPs