A desktop flash game to balance a ball on a stick. Implementation of Machine learning techniques to play the game autonomously and achieve highest possible score. An attempt to solve a real world balancing problem with the help of different machine learning algorithms in a simulated environment.
This branch has code for a supervised MLP model. The branch master_ga has the code for training the MLP with Genetic Algorithm.
The project is developed for a Linux OS that uses:
- OpenGL
- GLUT
- Bullet Physics Engine
- Keras ML Wrapper with Tensorflow
- C++ X11 library support
- C++ pthread