What is a “value function” in Markov Decision Processes?

a) The probability of transitioning between states

b) The expected return from a state under a particular policy

c) The time taken to reach a steady state

d) The total number of states in the process