FrozenBurning / RL-Mountain-Car-v0

Implementations of solutions to the both discrete and continuous mountain car problems.

Mountain Car v0 & v0-Continuous

Author: Zhaoxi Chen

Feature:

• Solved Mountain Car-v0 and Mountain Car-v0 Continuous problems
• Using Q-Learning, Sarsa, Exp-Sarsa
• Implement DQN as solution to continuous state space

Dependecies

Python 3.6+ gym

Usage

python scripts with "continuous_" prefix are solutions to Mountain Car-v0 continuous while "discrete_" prefix imply solutions to Mountain Car-v0

E.g. If you want to take Q-learning as solution to Mountain Car-v0, run as the following:

python3 discrete_qlearning.py

All of the scipts will save qtable as .npy and also plot reward curves.