Snake Game with Q-learning

This project implements a reinforcement learning agent based on Q-learning algorithm to play the classic Snake game. The agent learns to navigate the game environment, avoiding obstacles and collecting food to maximize its score.

Introduction

The project aims to demonstrate the application of reinforcement learning to a simple environment—the Snake game. The agent uses Q-learning, a model-free reinforcement learning algorithm, to learn how to play the game by maximizing its cumulative reward

How It Works

• Game Environment: The Snake game is implemented using Pygame, where the snake navigates a grid to collect food. The game ends when the snake collides with the walls or itself.
• Q-learning Agent: The agent observes the game state (snake and food positions) and takes actions (move up, down, left, right) based on a Q-table, which is updated through learning from experiences.
• Reward System: The agent receives positive rewards for collecting food and negative rewards for collisions, guiding it towards optimal behavior.

Training the Agent

The agent is trained over multiple episodes. During each episode:

• The agent selects actions based on an epsilon-greedy policy.
• The Q-table is updated using the Bellman equation.
• The exploration rate decays over time to balance exploration and exploitation.
• Training progress and total rewards are printed for each episode.

Evaluating the Agent

After training, the agent can be evaluated by running the game and observing its performance. The Q-table is used to select the best action for each game state, with no exploration.