Simple Solvers for MountainCar-v0 and MountainCarContinuous-v0 @ gym. Methods including Q-learning, SARSA, Expected-SARSA, DDPG and DQN.

• gym
• pytorch 1.3.1
• torchvision 0.4.2

Before run any script, please check out the parameters defined in the script and modify any of them as you please.

python TD.py

Adapted from REINFORCEMENT LEARNING (DQN) TUTORIAL in pytorch tutorials, which originally deals with CartPole Problem.

DQN method has not been run and tested.

python DQN.py

python inference.py

Adapted from Deep Reinforcement Learning Algorithms with PyTorch but rewritten in complete pytorch format, and redundant functions are removed.

python train_continuous.py