Dynamic Programming for Structured Continuous Markov Decision Problems

Zhengzhu Feng Richard Dearden Nicolas Meuleau Richard Washington

Published in: CoRR (2012)

Keyphrases

markov decision problems
dynamic programming
linear programming
state space
action space
optimal policy
partially observable

reinforcement learning
transition probabilities
decision theoretic
utility function
expected utility
markov decision processes
decision processes

infinite horizon
queueing networks
policy iteration
markov chain
decision problems
linear program

blocks world
real valued
average cost
dynamical systems
sufficient conditions
multi agent