Clash of Clans Gymnasium Environment

A custom OpenAI Gym environment and Pygame GUI for simulating Clash of Clans–style attack mechanics, combined with a Proximal Policy Optimization (PPO) agent (via Stable-Baselines3) to learn optimal attacking strategies.

---

Project Overview

• Environment: A Gym-compatible environment (WarzoneEnv) that models base layouts, troop deployment, and reward functions for RL.

• GUI: A Pygame + pygame_gui application (app.py) with screens for:

  • Base design
  • Troop selection
  • Attack simulation (where the agent plays)

• Agent: A PPO-based reinforcement learning agent implemented in model.py using Stable-Baselines3.

---

Dependeicies

• OpenAI Gym
• Stable-Baselines3
• Pygame
• pygame_gui

---

Installation

1. Clone this repo:

   git clone https://github.com/your-org/ClashGymEnv.git
   cd ClashGymEnv

2. (Optional) Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate      # macOS / Linux
   venv\Scripts\activate.bat   # Windows

3. Install dependencies:

   pip install -r requirements.txt

---

Training the Agent

1. Open model.py and configure your hyperparameters (learning rate, timesteps, etc.).

2. Run the training script:

   python model.py --train --timesteps 1000000

3. After training completes, a file ppo_model.zip (or similar) will be saved in the project root.

---

Running the GUI App

With a trained model in place, launch the Pygame application:

python app.py

1. Main Screen: Select your Town Hall level.
2. Base Design: Lay out defenses and buildings.
3. Troop Selection: Choose your deck composition.
4. Attack Simulation: Watch the PPO agent deploy troops and execute its learned strategy.

---

Adding New Algorithms

• Create a new agent file in agents/ (e.g., q_learning.py).

• Implement an Agent class with methods:

  • select_action(state)
  • train(env, episodes)
  • save(path) / load(path)

• In ui_attack_screen.py, import and instantiate your agent alongside the existing PPO agent.

---

LICENSE

See LICENSE