Balancing a pole with the (deep)mind

Application of Q-Learning and Double DQN to the CartPole problem. Final project for the course of Reinforcement Learning in MSc Robotics&AI Engineering @ University of Trieste, Dept. of Engineering and Architecture, Trieste.

Setup

To clone this repository, you need to have git installed. After that you can open a terminal window and run

git clone https://github.com/lorenzoelia/balancing_a_pole_with_the_deepmind

Move to the project folder

cd balancing_a_pole_with_the_deepmind

Now you need to install the dependencies. Make sure you have pip installed and then run

pip install -r requirements.txt

To run the program, do the following

cd src

and finally

python main.py

Directories

• src is the basis package containing all Python code. Inside there are:
  • Package baseline, which contains the baseline approach of random walk used for comparing Q-Learning and Double DQN effectiveness
  • Package dqn_learning_v0, which contains the classes necessary to train the DQN with Double Q-Learning. Folder models_v0 contains the trained DQN agents' weight
  • Package pole_test, to test the execution of the CartPole environment from Gym
  • Package q_learning, which contains the files for plain q-learning and testing