DocumentCode :
3129739
Title :
A Two-Timescale Simulation-Based Gradient Algorithm for Weighted Cost Markov Decision Processes
Author :
He, Ying ; Fu, Michael C. ; Marcus, Steven I.
Author_Institution :
Lister Hill National Center for Biomedical Communications, National Library of Medicine, Bethesda, MD 20894, USA. yihe@mail.nih.gov
fYear :
2005
fDate :
12-15 Dec. 2005
Firstpage :
8022
Lastpage :
8027
Abstract :
We develop a novel two-timescale simulation-based gradient algorithm for weighted cost Markov Decision Process (MDP) problems, illustrate the effectiveness of this algorithm by carrying out numerical experiments on a parking example, and compare the algorithm with two other algorithms in the literature.
Keywords :
Algorithm design and analysis; Approximation algorithms; Biomedical communication; Cost function; Helium; Infinite horizon; Libraries; Probability distribution; Stochastic processes; Stochastic systems;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Decision and Control, 2005 and 2005 European Control Conference. CDC-ECC '05. 44th IEEE Conference on
Print_ISBN :
0-7803-9567-0
Type :
conf
DOI :
10.1109/CDC.2005.1583460
Filename :
1583460
Link To Document :
بازگشت