In Section 1 we introduced an example with two states and two actions in which the reward function was \(r(A, 1)=4, r(B, 1)=3\), \(r(A, 2)=2, r(B, 2)=5\) and the transition matrices were as below. For a finite horizon stochastic dynamic programming problem with time horizon \(T=6\) and terminal reward \(R(A)=3, R(B)=5\), find the optimal policy.

Transcribed Image Text:

1/3 2/3 = (3/4 1/4) 14) and T = 1/5 4/5 1/4 3/4