POMCP enhancement for Hidden Semi-Markovian Mode MDP.
HS3MDP is based on POMCP 1.0 used in the NIPS 2010 paper Online Monte-Carlo Planning in Large POMDPs by David Silver and Joel Veness.
The original code can be found in this repository under the tag POMCP-1.0
.
Traffic, Elevator and Sailboat problems are modified versions of those proposed by Samuel Ping-Man Choi in his thesis Reinforcement Learning in Non-stationary Environments.
The presentation on this work, in the 8th joint NII-LIP6 workshop, can be found here.
- git
- automake, autoconf, etc.
- C++ Boost
git clone
this repository- run
autoreconf -i
- run
./configure
(possibly with--enable-assert
) - run
make
You will find the executable pomcp
in the src
directory.
Simply run pomcp --help
to see all possible parameters.