DMCA
Markov Decision Problems (1996)
| Venue: | U.C. Berkeley Randomized Algorithms Class Project |
| Citations: | 1 - 1 self |
BibTeX
@INPROCEEDINGS{Oza96markovdecision,
author = {Nikunj Oza},
title = {Markov Decision Problems},
booktitle = {U.C. Berkeley Randomized Algorithms Class Project},
year = {1996}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Markov Decision Problems (MDPs) are the foundation for many problems that are of interest to researchers in Artificial Intelligence and Operations Research. In this paper, we will review what is known about algorithms for solving MDPs as well as the complexity of solving MDPs in general. We will argue that, even though there are theoretically efficient algorithms for solving MDPs, these algorithms are inefficient enough in practice that efficient algorithms for finding exact and approximate solutions to MDPs are still needed. We will review some aspects of MDPs that we hope can potentially be exploited by these currently unknown algorithms. 1. Introduction A Markov Chain is a stochastic process that satisfies the Markov Property: what the next state will be depends only on the current state and not on any past states. A Markov Decision Problem is a Markov Chain with the addition of an action model and a performance criterion. Actions change the state transition matrix, giving some co...
