This repository contains a collection of algorithms used to solve the Cartpole environment on OpenAI Gym.
Cartpole balancing is a control problem where the goal is to balance a pole upright on the top of a cart by applying the appropriate forces left and right.
States:
- The position of the cart
- The velocity of the cart
- The angle of the pole
- The angular velocity of the pole
Actions:
- Apply a force of +1 (move the cart right)
- Apply a force of -1 (move the cart left)
Three algorithms were used for this project:
- A Deep Q-Network with Experience Replay
- A Genetic Algorithm
- A PID Controller with Gradient Descent
- Python3 (Note: ensure you have python mapped to the python3 environment variable)
- NumPy
- Scikit-learn (Note: ensure 0.22.1 is the version of Scikit-learn being used)
- Matplotlib (Note: ensure that your operating system is able to use the TKAgg function)
- Keras (Note: ensure you are using a TensorFlow backend and 2.4.3 is the version of Keras used)
- TensorFlow (Note: ensure 2.4.1 is the version of TensorFlow being used)
- Pillow
- pip (Note: ensure you have pip mapped to the pip3 environment variable)
To install the project dependencies run the command "pip3 install -r requirements.txt"
The following steps detail how to run the project:
- Use the cd command in order to navigate to the "algorithms" directory
- Run python Plotter.py to see the results of all the algorithms and Baseline
- Run python Baseline.py to see the results of the Baseline alone
- Run python CartpoleDQN.py to see the results of the DQN alone
- Run python Genetic.py to see the results of the Genetic Algorithm alone
- Run python PIDController.py to see the results of the PID Controller alone
MSc Computer Science - Intelligent Systems Module: CS7IS2 - Artificial Intelligence
Name: Claire Farrell Student Number: 1619148
Name: Daniel Farrell Student Number: 18315021
Name: Joshua Cassidy Student Number: 20300057
Name: Matteo Bresciani Student Number: 20309566