maze ml

I'll write a more detailed README later.

The project consists of multiple Q-Networks that feed each other information as they get closer to the escape of the maze. The maze is randomly generated to assure that the Q-Networks can be accurate in multiple situations. Parameters can be changed like maze size, agent hyperparameters, etc.