124
click to vote
CDC
2010
IEEE
14 years 9 months ago
2010
IEEE
We consider the classical finite-state discounted Markovian decision problem, and we introduce a new policy iteration-like algorithm for finding the optimal state costs or Q-facto...