Multilinear and integer programming for markov decision processes with imprecise probabilities

Ricardo Shirota Filho; Fabio Gagliardi Cozman; Felipe Werndl Trevizan; Cassio Polpo De Campos; Leliane Nunes De Barros

Multilinear and integer programming for markov decision processes with imprecise probabilities

Ricardo Shirota Filho, Fabio Gagliardi Cozman, Felipe Werndl Trevizan, Cassio Polpo De Campos, Leliane Nunes De Barros

Research output: Contribution to conference › Conference Paper

12 Scopus citations

Abstract

Markov Decision Processes (MDPs) are extensively used to encode sequences of decisions with probabilistic effects. Markov Decision Processes with Imprecise Probabilities (MDPIPs) encode sequences of decisions whose effects are modeled using sets of probability distributions. In this paper we examine the computation of Γ-maximin policies for MDPIPs using multilinear and integer programming. We discuss the application of our algorithms to "factored" models and to a recent proposal, Markov Decision Processes with Set-valued Transitions (MDPSTs), that unifies the fields of probabilistic and "nondeterministic" planning in artificial intelligence research. Copyright © 2007 by SIPTA.

Original language	American English
Pages	395-404
Number of pages	10
State	Published - 1 Dec 2007
Externally published	Yes
Event	ISIPTA 2007 - Proceedings of the 5th International Symposium on Imprecise Probability: Theories and Applications - Duration: 1 Dec 2007 → …

Conference

Conference	ISIPTA 2007 - Proceedings of the 5th International Symposium on Imprecise Probability: Theories and Applications
Period	1/12/07 → …

Cite this

@conference{da680ed729e448239ccb188528c84833,

title = "Multilinear and integer programming for markov decision processes with imprecise probabilities",

abstract = "Markov Decision Processes (MDPs) are extensively used to encode sequences of decisions with probabilistic effects. Markov Decision Processes with Imprecise Probabilities (MDPIPs) encode sequences of decisions whose effects are modeled using sets of probability distributions. In this paper we examine the computation of Γ-maximin policies for MDPIPs using multilinear and integer programming. We discuss the application of our algorithms to {"}factored{"} models and to a recent proposal, Markov Decision Processes with Set-valued Transitions (MDPSTs), that unifies the fields of probabilistic and {"}nondeterministic{"} planning in artificial intelligence research. Copyright {\textcopyright} 2007 by SIPTA.",

author = "Filho, {Ricardo Shirota} and Cozman, {Fabio Gagliardi} and Trevizan, {Felipe Werndl} and {De Campos}, {Cassio Polpo} and {De Barros}, {Leliane Nunes}",

year = "2007",

month = dec,

day = "1",

language = "American English",

pages = "395--404",

note = "ISIPTA 2007 - Proceedings of the 5th International Symposium on Imprecise Probability: Theories and Applications ; Conference date: 01-12-2007",

}

Multilinear and integer programming for markov decision processes with imprecise probabilities. / Filho, Ricardo Shirota; Cozman, Fabio Gagliardi; Trevizan, Felipe Werndl; De Campos, Cassio Polpo; De Barros, Leliane Nunes.

2007. 395-404 Paper presented at ISIPTA 2007 - Proceedings of the 5th International Symposium on Imprecise Probability: Theories and Applications.

Research output: Contribution to conference › Conference Paper

TY - CONF

T1 - Multilinear and integer programming for markov decision processes with imprecise probabilities

AU - Filho, Ricardo Shirota

AU - Cozman, Fabio Gagliardi

AU - Trevizan, Felipe Werndl

AU - De Campos, Cassio Polpo

AU - De Barros, Leliane Nunes

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Markov Decision Processes (MDPs) are extensively used to encode sequences of decisions with probabilistic effects. Markov Decision Processes with Imprecise Probabilities (MDPIPs) encode sequences of decisions whose effects are modeled using sets of probability distributions. In this paper we examine the computation of Γ-maximin policies for MDPIPs using multilinear and integer programming. We discuss the application of our algorithms to "factored" models and to a recent proposal, Markov Decision Processes with Set-valued Transitions (MDPSTs), that unifies the fields of probabilistic and "nondeterministic" planning in artificial intelligence research. Copyright © 2007 by SIPTA.

AB - Markov Decision Processes (MDPs) are extensively used to encode sequences of decisions with probabilistic effects. Markov Decision Processes with Imprecise Probabilities (MDPIPs) encode sequences of decisions whose effects are modeled using sets of probability distributions. In this paper we examine the computation of Γ-maximin policies for MDPIPs using multilinear and integer programming. We discuss the application of our algorithms to "factored" models and to a recent proposal, Markov Decision Processes with Set-valued Transitions (MDPSTs), that unifies the fields of probabilistic and "nondeterministic" planning in artificial intelligence research. Copyright © 2007 by SIPTA.

UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79957825456&origin=inward

UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=79957825456&origin=inward

M3 - Conference Paper

SP - 395

EP - 404

T2 - ISIPTA 2007 - Proceedings of the 5th International Symposium on Imprecise Probability: Theories and Applications

Y2 - 1 December 2007

ER -

Multilinear and integer programming for markov decision processes with imprecise probabilities

Abstract

Conference

Other files and links

Fingerprint

Cite this