Reinforcement Learning Snake Game - DQN

By Sudharsan Ananth

Table of Contents

1. Introduction
2. Dependencies
3. Prerequisites
4. Run the game
5. License

Introduction

In this project I have created a DQN agent to train in a snake game environment. The objective of the snake is to collide(eat) all the food that is placed randomly in the environment.

Untrained agent.

Trained agent.

Dependencies

This project is built with the below given major frameworks and libraries. The code is primarily based on python. And the environment is created using Anaconda.

• Python
• Anaconda
• pytorch
• matplotlib
• Pygame

Prerequisites

1. Python 3.7 (skip if downloading conda)
2. Conda (Either miniconda or Ananconda)

Run the game

Simply clone the repo cd into the right directory and run agent using the below commands. Step-by-Step instructions given below. Simply change the directory and run agent.py from the directory RL_car_game. You will be able to see the agent training and getting better in minutes.

1. Clone the repository using

   git clone https://github.com/sudharsan-007/RL-DQN-Snake-Game.git

2. cd into the directory RL-DQN-Snake-Game

   cd RL-DQN-Snake-Game

3. Recommended: create a conda environment

   # We require python>=3.7
   conda create -n rl_snake_game python=3.7 numpy matplotlib 
   conda activate rl_snake_game

4. Install pygame

   pip install pygame
   pip install ipython
   pip install matplotlib

5. Install pyTorch (CPU verison). Please refer pytorch website to get right version for GPU.

   # https://pytorch.org/get-started/locally/
   conda install pytorch torchvision torchaudio cpuonly -c pytorch

6. If the above code does not work for you try this(skip this step 5 worked).

   pip3 install torch torchvision torchaudio

7. Run agent.py from this directory and from inside this environment

   python agent.py

8. To run the environment without Reinforcement Agent and the agent controllable by WASD keys

   python snake_game.py

License

Distributed under the MIT License. See LICENSE.txt for more information.

(back to top)