Fitted Natural Actor-Critic: A New Algorithm for Continuous State-Action MDPs.

Francisco S. Melo Manuel Lopes

Published in: ECML/PKDD (2) (2008)

Keyphrases

dynamic programming
natural actor critic
learning algorithm
reinforcement learning
objective function
np hard
robot arm
optimal solution
support vector machine svm
path planning
markov decision processes
convergence rate
feature space
mobile robot
linear programming
kernel methods
state action