Title :
Adaptive computation of optimal nonrandomized policies in constrained average-reward MDPs
Author :
Feinberg, Eugene A.
Author_Institution :
Dept. of Appl. Math. & Stat., Stony Brook Univ., Stony Brook, NY
fDate :
March 30 2009-April 2 2009
Abstract :
This paper deals with computation of optimal nonrandomized nonstationary policies and mixed stationary policies for average-reward Markov decision processes with multiple criteria and constraints. We consider problems with finite state and action sets satisfying the unichain condition. The described procedure for computing optimal nonrandomized policies can also be used for adaptive control problems.
Keywords :
Markov processes; action sets; adaptive computation; adaptive control problems; average-reward Markov decision processes; constrained average-reward MDP; finite state; mixed stationary policies; optimal nonrandomized nonstationary policies; unichain condition; Adaptive control; Constraint theory; Frequency; Linear programming; Mathematics; Probability distribution; State-space methods; Statistics; Vectors;
Conference_Titel :
Adaptive Dynamic Programming and Reinforcement Learning, 2009. ADPRL '09. IEEE Symposium on
Conference_Location :
Nashville, TN
Print_ISBN :
978-1-4244-2761-1
DOI :
10.1109/ADPRL.2009.4927531
