Solving LunarLander-v2 with Tabular, DQN and REINFORCE using TensorFlow

• Run the notebooks on Google Colab is recommended
• IMPORTANT: please change the notebook_path in the initialization section of all notebooks when running on Google Colab.

• custom_animation.py contains an implemented package for video demonstration on episodes.
• custom_tdcontrol.py contains implementations of SARSA and tabular Q-learning algorithms.
• Demo.ipynb contains video demo and some discussions on the algorithms.
• DQN.ipynb, REINFORCE.ipynb and TDControl.ipynb train their corresponding algorithms.
• models folder contains the trained TensorFlow models of DQN and REINFORCE.