PongEvolution

Neural nets learn to play pong using the genetic algorithm.

How to use

Edit the values in config.py to your liking:

• SCREENSIZE (tuple): The size of the screen (w,h)
• PADDLESIZE (tuple): The default paddle dimensions
• PLAYERSPEED (number): The maximum speed the paddles can travel at
• HIDDENNODES (int): The number of hidden nodes each player's neural net contains
• BALLRADIUS (number): The radius of the balls
• DEFAULT_BALLSPEED (number): The base speed the balls start at
• DIFFICULTY_RATE (number): The rate at which the ball speed increases each frame
• POPULATION_SIZE (number): The size of the population
• MUTATION_RATE (number): The rate at which the organisms mutate
• FPS_CAP (number): The maximum FPS to run at. -1 for no max

Run main.py

• Use alt to show the neural net of the winner of the previous generation
• Use shift to kill every organism in the current generation
• Hold w to only show the previous generation's best player