Proximal Policy Optimization (PPO) for Farama Foundation Highway Environment

This repository contains an implementation of Proximal Policy Optimization (PPO) for the Farama Foundation Highway Environment. PPO is a reinforcement learning algorithm used to train agents to make sequential decisions, and it is applied to solve the specific task of navigating the Farama Foundation Highway Environment.

Table of Contents

• Introduction
• Installation
• Usage
• Contributing
• License

Introduction

The Farama Foundation Highway Environment is a environment for training reinforcement learning agents. This repository provides a PPO implementation for training an agent to navigate this environment effectively. The implementation includes all the necessary components for training and evaluating the agent's performance.

Installation

1. Clone this repository to your local machine:

   git clone https://github.com/HasarinduPerera/ppo-highway-env.git

2. Install the required dependencies:

   cd ppo-highway-env
   pip install -r requirements.txt

Usage

1. Train the PPO agent:

   python main.py

   You can modify hyperparameters, training settings, and network architectures to suit your specific needs.

2. Evaluate the trained agent:

   python inference.py -mp /path/to/pre-trained-model -i 10

   The evaluation script will load the trained model and test the agent's performance in the environment.

   • -mp or --model-path: Path to the pre-trained model (required).
   • -i or --inference-iterations: Number of inference iterations (default: 10).

3. Adjust and fine-tune the code as needed for your specific use case.

Contributing

Contributions are welcome! If you'd like to contribute to this project, please follow these steps:

1. Fork the repository
2. Create a new branch for your feature or bug fix
3. Make your changes and commit them with descriptive messages
4. Push your branch to your fork
5. Create a pull request to this repository's main branch

License

This project is licensed under the GNU General Public License v3.0 - see the LICENSE file for details.