An implementation of Tetris where an AI player is trained using genetic algorithms to develop optimal playing strategies.

• Features
• Implementation Details
• Requirements
• Installation

• Complete Tetris Implementation: Fully functional Tetris game with all standard mechanics
• AI Player: Genetic algorithm-based AI that learns to play Tetris
• Multiple Evaluation Heuristics: 9 different board metrics used for move evaluation
• Training System: Configurable genetic algorithm training pipeline
• Visualization: Real-time display of the AI's gameplay
• Logging: Detailed logs of the evolutionary process and performance metrics

The AI uses a genetic algorithm to evolve optimal weights for evaluating Tetris moves based on several heuristics:

1. Board Metrics:

   • Maximum height
   • Total height
   • Bumpiness
   • Holes
   • Future risk

2. Genetic Algorithm:

   • Population size: 12 chromosomes
   • Generations: 10+
   • Selection: Top 50% elitism
   • Crossover: Uniform crossover
   • Mutation: Gaussian noise with 10% rate

3. Move Evaluation:

   • Tests all possible rotations and positions
   • Scores each potential move using weighted heuristics
   • Selects the highest scoring valid move

• Python 3.7+
• Pygame 2.0+

1. Clone the repository:

   git clone https://github.com/mariemomar/Tetris_Game.git
   cd Tetris_Game