A Linearly Relaxed Approximate Linear Program for Markov Decision Processes
Systems and Control
2017-04-11 v1
Abstract
Approximate linear programming (ALP) and its variants have been widely applied to Markov Decision Processes (MDPs) with a large number of states. A serious limitation of ALP is that it has an intractable number of constraints, as a result of which constraint approximations are of interest. In this paper, we define a linearly relaxed approximation linear program (LRALP) that has a tractable number of constraints, obtained as positive linear combinations of the original constraints of the ALP. The main contribution is a novel performance bound for LRALP.
Cite
@article{arxiv.1704.02544,
title = {A Linearly Relaxed Approximate Linear Program for Markov Decision Processes},
author = {Chandrashekar Lakshminarayanan and Shalabh Bhatnagar and Csaba Szepesvari},
journal= {arXiv preprint arXiv:1704.02544},
year = {2017}
}
Comments
23 pages, 2 figures, submitted to IEEE TAC