This repo contains the code for a simple, model-free reinforcement learning implementation of the cliff walk from Sutton and Barto. Both SARSA and Qlearning are implemented.
In MATLAB, run parameters.m
Here is the final path for the SARSA agent:
Here are the updates to the Q values for the SARSA agent:
Here is the final path for the Q learning agent:
Here are the updates to the Q values for the Q learning agent:
Update the image saver and gif maker file names.