Maze Runner Project

Overview

The Maze Runner project is a Python-based application to solve dynamic mazes using artificial intelligence techniques. It employs Q-learning, artificial life for dynamic wall updates, and rule-based systems for moving obstacles (grievers) to challenge the maze-solving capabilities of an agent.

---

Files and Their Purpose

1. mazeParser.py:

   • Purpose:
     • Handles loading, parsing, and visualizing mazes.
     • Manages dynamic wall changes during the simulation.
   • Key Functions:
     • load_maze: Loads a maze from a CSV file.
     • load_dynamic_walls: Loads dynamic wall configurations.
     • visualize_maze: Visualizes the static maze.
     • apply_dynamic_wall_changes: Updates maze structure based on dynamic walls.

2. qlearning.py:

   • Purpose: Implements the Q-learning algorithm to train the agent to solve the maze.
   • Key Classes:
     • MazeRunner:
       • Initializes and trains a Q-table for decision-making.
       • Simulates agent movements and computes rewards based on maze dynamics.
     • Key methods include:
       • train: Trains the agent using Q-learning with dynamic wall and griever interactions.
       • find_path: Extracts the optimal path after training.

3. main.py:

   • Purpose: Acts as the entry point for the application.
   • Key Features:
     • Provides a GUI for selecting input files (maze and dynamic walls).
     • Runs the simulation by integrating maze visualization, Q-learning, and live updates.
     • Combines and executes functionalities from other modules.

4. grievers.py:

   • Purpose: Handles the behavior of moving obstacles (grievers) in the maze with rule based systems.
   • Key Function:
     • update_griever_positions: Dynamically updates griever positions while ensuring valid moves within the maze.

---

Setting Up the Environment (macOS)

1. Prerequisites:

   • Python 3.8 or newer.
   • pip for package management.

2. Install Required Libraries: Open a terminal and run the following commands

   • brew install [email protected]
     • To install tkinter
     • Run python -m tkinter to check if tkinter is installed
   • pip install -r requirements.txt
     • To install all of the dependencies for the application

3. Running the Application:

   • Open the terminal and navigate to the directory containing the project files:

     cd /path/to/main.py

   • Start the application:

     python main.py

   • Use the GUI to select the maze and dynamic walls input files.

---

Artificial Intelligence Techniques Used

1. Q-Learning:

   • A model-free reinforcement learning algorithm.
   • Updates a Q-table with state-action pairs to learn an optimal policy.
   • Exploits a reward system to guide the agent towards the goal while avoiding grievers and walls.

2. Artificial Life:

   • Simulates dynamic walls that change the maze structure during the simulation based on triggers.
   • Creates evolving and adaptive behaviors in the environment.

3. Rule-Based Systems:

   • Governs the movement of grievers using predefined rules.
   • Ensures grievers move in a predictable yet dynamic fashion within the maze.

4. Exploration vs. Exploitation:

   • Implements an epsilon-greedy strategy to balance exploring new paths and exploiting known rewards.

5. Pathfinding Assistance:

   • Employs heuristic rewards to encourage movements closer to the goal.

---

This project demonstrates the integration of reinforcement learning with artificial life and rule-based systems, making it suitable for exploring AI techniques in adaptive environments. Enjoy the challenge of solving dynamic mazes!