Consider an agent starting in a room a in which it can take two possible actions: to leave the room (action "l") or to stay (action "s"). if it leaves a, the agent moves to room b, which is a terminal state (no more actions can be taken). the outcomes of the actions are uncertain, so that when executing action l (or action s), there is some probability that the agent will leave a (or stay in a). we assume that the reward in entering state b is r(b) = +1 and the reward for being in state a is r(a) = -0.1. (a) draw the (very simple) diagram corresponding to this mdp. answer by inspection of the diagram: what is the optimal policy? (b) assume that the agent knows neither the world (transition probabilities) nor the utilities of the states. assume that the agent, for some reason, happens to follow the optimal policy. the rewards received at states a and b are the same as described above.. in the process of executing this policy, the agent execute four trials and, in each trial, it stops after reaching state b. the following state sequences are recorded during the trials: aaab, aab, ab, ab. what is the estimate of t., what is the estimate of u(a), assuming a discount factor of = 0.5?