Singularly perturbed linear programs and Markov decision processes

Konstantin Avrachenkov; Jerzy Filar; Vladimir Gaitsgory; Andrew Stillman

doi:10.1016/j.orl.2016.02.005

Singularly perturbed linear programs and Markov decision processes

Konstantin Avrachenkov, Jerzy Filar, Vladimir Gaitsgory, Andrew Stillman

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

Abstract

Linear programming formulations for the discounted and long-run average MDPs have evolved along separate trajectories. In 2006, E. Altman conjectured that the two linear programming formulations of discounted and long-run average MDPs are, most likely, a manifestation of general properties of singularly perturbed linear programs. In this note we demonstrate that this is, indeed, the case.

Original language	English
Pages (from-to)	297-301
Number of pages	5
Journal	Operations Research Letters
Volume	44
Issue number	3
DOIs	https://doi.org/10.1016/j.orl.2016.02.005
Publication status	Published - 1 May 2016

Keywords

Discounted MDPs
Limiting linear program
Long-run average MDPs
Markov Decision Processes (MDPs)
Singularly perturbed linear programs

Access to Document

10.1016/j.orl.2016.02.005

Cite this

@article{fb7ede3865a3427fab12fe34242d0fda,

title = "Singularly perturbed linear programs and Markov decision processes",

abstract = "Linear programming formulations for the discounted and long-run average MDPs have evolved along separate trajectories. In 2006, E. Altman conjectured that the two linear programming formulations of discounted and long-run average MDPs are, most likely, a manifestation of general properties of singularly perturbed linear programs. In this note we demonstrate that this is, indeed, the case.",

keywords = "Discounted MDPs, Limiting linear program, Long-run average MDPs, Markov Decision Processes (MDPs), Singularly perturbed linear programs",

author = "Konstantin Avrachenkov and Jerzy Filar and Vladimir Gaitsgory and Andrew Stillman",

year = "2016",

month = may,

day = "1",

doi = "10.1016/j.orl.2016.02.005",

language = "English",

volume = "44",

pages = "297--301",

journal = "Operations Research Letters",

issn = "0167-6377",

publisher = "Elsevier",

number = "3",

}

TY - JOUR

T1 - Singularly perturbed linear programs and Markov decision processes

AU - Avrachenkov, Konstantin

AU - Filar, Jerzy

AU - Gaitsgory, Vladimir

AU - Stillman, Andrew

PY - 2016/5/1

Y1 - 2016/5/1

N2 - Linear programming formulations for the discounted and long-run average MDPs have evolved along separate trajectories. In 2006, E. Altman conjectured that the two linear programming formulations of discounted and long-run average MDPs are, most likely, a manifestation of general properties of singularly perturbed linear programs. In this note we demonstrate that this is, indeed, the case.

AB - Linear programming formulations for the discounted and long-run average MDPs have evolved along separate trajectories. In 2006, E. Altman conjectured that the two linear programming formulations of discounted and long-run average MDPs are, most likely, a manifestation of general properties of singularly perturbed linear programs. In this note we demonstrate that this is, indeed, the case.

KW - Discounted MDPs

KW - Limiting linear program

KW - Long-run average MDPs

KW - Markov Decision Processes (MDPs)

KW - Singularly perturbed linear programs

UR - http://www.scopus.com/inward/record.url?scp=84959360195&partnerID=8YFLogxK

U2 - 10.1016/j.orl.2016.02.005

DO - 10.1016/j.orl.2016.02.005

M3 - Article

SN - 0167-6377

VL - 44

SP - 297

EP - 301

JO - Operations Research Letters

JF - Operations Research Letters

IS - 3

ER -

Singularly perturbed linear programs and Markov decision processes

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this