TD-Lambda-Online-Learning

In this project we are going to implement td lambda learning. There are two approaches to implement the td lambda learning: 1. Forward or the theoretical view (Offline learning) 2.Backward or practical view (Online Learning) Here we used simple random walk envoironment to see the performance of td lambda with different lambda numbers and different alphas. Since the actual value of the states are available for us, we used RMS to calculate the error rate. In the sutton's book it is mentioned that the backward view is equivalent to the forward view. So this implementation is enough in most of the cases.