Singularly Perturbed Markov Decision Processes: A Multiresolution Algorithm.

Chin Pang Ho Panos Parpas

Published in: SIAM J. Control. Optim. (2014)

Keyphrases

markov decision processes
dynamic programming
multiresolution
learning algorithm
policy iteration
average reward
computational complexity
search space
state space
model based reinforcement learning
np hard
reinforcement learning
linear programming
optimal policy
optimal solution
finite horizon
objective function