We present new numerical algorithms and bounds for the infinite horizon,
discrete stage, finite state and action Markov decision process with
imprecise transition probabilities. We assume that the transition
probability mass vector for each state and action is described by
a finite number of linear inequalities. This model of imprecision
appears to be well suited for describing statistically determined
confidence limits and/or natural language statements of likelihood.
The numerical procedures for calculating an optimal max-mm strategy
are based on successive approximations, reward revision, and modified
policy iteration. The bounds that are determined are at least as
tight as currently available bounds for the case where the transition
probabilities are precise.
%0 Journal Article
%1 White:1994
%A III, Chelsea C. White
%A Eldeib, Hany K.
%D 1994
%J Operations Research
%K *ALGORITHMS *DECISION *MARKOV -- ATOMIC LINEAR MARKOV Mathematical making models probabilities processes spectrum systems transition
%P 739-749
%T Markov decision processes with imprecise transition probabilities
%V 42
%X We present new numerical algorithms and bounds for the infinite horizon,
discrete stage, finite state and action Markov decision process with
imprecise transition probabilities. We assume that the transition
probability mass vector for each state and action is described by
a finite number of linear inequalities. This model of imprecision
appears to be well suited for describing statistically determined
confidence limits and/or natural language statements of likelihood.
The numerical procedures for calculating an optimal max-mm strategy
are based on successive approximations, reward revision, and modified
policy iteration. The bounds that are determined are at least as
tight as currently available bounds for the case where the transition
probabilities are precise.
@article{White:1994,
abstract = {We present new numerical algorithms and bounds for the infinite horizon,
discrete stage, finite state and action Markov decision process with
imprecise transition probabilities. We assume that the transition
probability mass vector for each state and action is described by
a finite number of linear inequalities. This model of imprecision
appears to be well suited for describing statistically determined
confidence limits and/or natural language statements of likelihood.
The numerical procedures for calculating an optimal max-mm strategy
are based on successive approximations, reward revision, and modified
policy iteration. The bounds that are determined are at least as
tight as currently available bounds for the case where the transition
probabilities are precise.},
added-at = {2009-06-26T15:25:19.000+0200},
author = {III, Chelsea C. White and Eldeib, Hany K.},
biburl = {https://www.bibsonomy.org/bibtex/2a9da393235f2fc5d77da259f341492f6/butz},
description = {diverse cognitive systems bib},
interhash = {6278136b37b5031d98f7d562c82a43d6},
intrahash = {a9da393235f2fc5d77da259f341492f6},
journal = {Operations Research},
keywords = {*ALGORITHMS *DECISION *MARKOV -- ATOMIC LINEAR MARKOV Mathematical making models probabilities processes spectrum systems transition},
owner = {butz},
pages = {739-749},
timestamp = {2009-06-26T15:26:01.000+0200},
title = {Markov decision processes with imprecise transition probabilities},
volume = 42,
year = 1994
}