Reinforcement Learning for an Inverted Pendulum

In this repository control of inverse pendel has been implemented

Q-Learning with GYM cartpole:

1. Install the requirements.txt with ' pip install requirments.txt '

2. the settings of the training can be changed in config.py

3. navigate to /Q-learning for training run: 'python main.py -t' after this a .npy file with the Q-table will be saved under weights

4. The Q-table can be used to visualize the control via 'python main.py'

The space-state of the cartpole was implemented in this repository: https://github.com/openai/gym/blob/master/gym/envs/classic_control/cartpole.py

TODO:

• Impalement DQN
• Impalement DDQN

Authors:

• Danikhani

License

This project is licensed under the MIT License