bouncingballs

Simple 2D game solved using Deep Q-Learning

Hit all the Green Balls while avoiding the Red ones. The balls coordinates and velocities are randomly generated at the start of each game.

3 scripts can be run:

• play.py: Play by moving agent using the KEYPAD keys

• train.py: Create Q-Learner using Tensorflow and learn by simulation

• simulate.py: Launch a game simulation using the calibrated agent

To train the model, you need to create a folder called "saved" in the project Root directory