Basic MonteCarlo search algorithms as described in page 673
http://www.jair.org/papers/paper3484.html
### Requirements
git clone https://github.com/openai/gym
cd gym
pip install -e . # minimal install
QValueFunction uses linear approximation of the Q function applied to Mountain Car problem
- mc_search.py (please run this file)
- QValueFunction.py
- util.py
QValueFunction uses Theanets Regressor as function approximator applied to Mountain Car problem
- mc_search_theanets.py (please run this file)
- QNetwork_theanets.py
- util.py
Reference benchmark using random choice on every step.
- Pendulum_baseline.py (please run this file)
QValueFunction uses Theanets Regressor as function approximator applied to Pendulum problem
- Pend_mc_search_theanets.py (please run this file)
- Pend_QNetwork_theanets.py
- util.py