A smartcab based on reinforcement learning.
Requirements:
This project requires pygame library to run the simulations for the smartcab.
About:
Q-learning has been used with an optimal decay function for epsilon to train the smartcab to create a safe and reliable smartcab. Main implementation is done in 'smartcab/agent.py' with helper functions provided in 'smartcab/environment.py' , 'smartcab/planner.py' and 'smartcab/simulator.py' for setting up the environment for the smartcab and creating simulations of the cab and the environment.