Example 2 - state aliased grid-world

Consider the following state aliased grid-world:

In the previous diagram, we see there are eight states, where two states with reward -10 are bad states to avoid and there's one state with reward 10 is a good state and also the goal state. However, our major concerns are the gray-shaded states, where the agent can't differentiate between these two states as they share the same features. 

Say, the feature for any state s be  and actions are North, East, West, and South. 

Considering these features, let's compare ...

Get Reinforcement Learning with TensorFlow now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.