Maze Solving with Q-learning