1 / 19

MDP Reinforcement Learning

MDP Reinforcement Learning. Markov Decision Process. “Should you give money to charity?”. “Should you give money to charity?”. “Would you contribute?”. “Would you contribute?”. $. Charity MDP. State space : 3 states

maxine
Download Presentation

MDP Reinforcement Learning

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. MDP Reinforcement Learning

  2. Markov Decision Process “Should you give money to charity?” “Should you give money to charity?” “Would you contribute?” “Would you contribute?” $

  3. Charity MDP • State space : 3 states • Actions : “Should you give money to charity” ,“Would you contribute” • Observations : knowledge of current state • Rewards : in final state, positive reward proportional to amount of money gathered

  4. So how can we raise the most money (maximize the reward)? • I.e. What is the best policy? • Policy : optimal action for each state

  5. Lecture Outline • Computing the Value Function • Finding the Optimal Policy • Computing the Value Function in an Online Environment

  6. Useful definitions Define:  to be a policy (j) : the action to take in j R(j) the reward from a certain state f(j,) : the next state, starting from state j and performing action 

  7. Computing The Value Function • When the reward is known, we can compute the value function for a particular policy • V(j), the value function : Expected reward for being in state j, and following a certain policy 

  8. Calculating V(j) • Set V0 (j) = 0, for all j • For i = 1 to Max_i • Vi (j) = R(j) +  V(i-1) (f(j, (j))) • = the discount rate, measures how much future rewards can propagate to previous states Above formula depends on the rewards being known

  9. Value Fn forthe Charity MDP • Fixing  at .5, and two policies, one which asks both • questions, and the other cuts to the chase • What is V3 if : • Assume that the reward is constant at the final state • (everyone gives the same amount of money) • 2. Assume that if you ask if one should give to charity, the • reward is 10 times higher.

  10. Given the value function, how can we find the policy which maximizes the rewards?

  11. Policy Iteration • Set 0 to be an arbitrary policy • Set i to 0 • Compute Vi(j) for all states j • Compute (i+1)(j) = argmax  Vi(f(j,)) • If (i+1) = i stop, otherwise i++ and back to step 3 What would this for the charity MDP for the two cases?

  12. Lecture Outline Computing the Value Function Finding the Optimal Policy Computing the Value Function in an Online Environment

  13. MDP Learning • So, the rewards are known, we can calculate the optimal policy using policy iteration. • But what happens in the case where we don’t know the rewards?

  14. Lecture Outline Computing the Value Function Finding the Optimal Policy Computing the Value Function in an Online Environment

  15. Deterministic vs. Stochastic Update Deterministic : Vi (j) = R(j) +  V(i-1) (f(j, (j))) Stochastic : V(n) = (1 - ) V(n) + [r + V(n’)] • Difference in that stochastic version averages over all visits to the state

  16. MDP extensions • Probabilistic state transitions • How should you calculate the value function for the first state now? “Would you like to contribute?” .8 ““Would you like to contribute?” .2 Mad Happy +10 -10

  17. Probabilistic Transitions • Online computation strategy works the same even when state transitions are unknown • Works in the case when you don’t know what the transitions are

  18. Online V(j) Computation • For each j initialize V(j) = 0 • Set n = initial state • Set r = reward in state n • Let n’ = f(n, (n)) • V(n) = (1 - ) V(n) + [r + V(n’)] • n = n’, and back to step 3

  19. 1-step Q-learning • Initialize Q(n,a) arbitraily • Select  as policy • n = initial state, r = reward, a = (n) • Q(n,a) = (1 - ) Q(n,a) + [r +  maxa’Q(n’,a’)] • n = n’, and back to step 3

More Related