Markov decision problems with unbounded transition rates under discounted-cost performance criteria

Baoqun Yin; Yanjie Li; Guiping Dai; Hu Zhang; Hongsheng Xi

Markov decision problems with unbounded transition rates under discounted-cost performance criteria

Baoqun Yin^*
, Yanjie Li
, Guiping Dai
, Hu Zhang
, Hongsheng Xi

^*Corresponding author for this work

University of Science and Technology of China

Research output: Contribution to conference › Paper › peer-review

Abstract

We discuss the problems of discounted-eost performance optimization for a class of countable Markov control processes (CMCP) with (possibly) unbounded transition rates. The discounted Poisson equation is proposed for a CMCP, from which α -potential is defined. The optimality equation based on the α -potential is derived under the conditions weaker than those previously known.

Original language	English
Pages	357-359
Number of pages	3
State	Published - 2004
Externally published	Yes
Event	WCICA 2004 - Fifth World Congress on Intelligent Control and Automation, Conference Proceedings - Hangzhou, China Duration: 15 Jun 2004 → 19 Jun 2004

Conference

Conference	WCICA 2004 - Fifth World Congress on Intelligent Control and Automation, Conference Proceedings
Country/Territory	China
City	Hangzhou
Period	15/06/04 → 19/06/04

Keywords

Countable Markov control processes
Discounted Poisson equations
Discounted-cost performance criteria
Unbounded transition rates
α -potential

Cite this

@conference{621a463496554eb580e5a01d5292228c,

title = "Markov decision problems with unbounded transition rates under discounted-cost performance criteria",

abstract = "We discuss the problems of discounted-eost performance optimization for a class of countable Markov control processes (CMCP) with (possibly) unbounded transition rates. The discounted Poisson equation is proposed for a CMCP, from which α -potential is defined. The optimality equation based on the α -potential is derived under the conditions weaker than those previously known.",

keywords = "Countable Markov control processes, Discounted Poisson equations, Discounted-cost performance criteria, Unbounded transition rates, α -potential",

author = "Baoqun Yin and Yanjie Li and Guiping Dai and Hu Zhang and Hongsheng Xi",

year = "2004",

language = "英语",

pages = "357--359",

note = "WCICA 2004 - Fifth World Congress on Intelligent Control and Automation, Conference Proceedings ; Conference date: 15-06-2004 Through 19-06-2004",

}

TY - CONF

T1 - Markov decision problems with unbounded transition rates under discounted-cost performance criteria

AU - Yin, Baoqun

AU - Li, Yanjie

AU - Dai, Guiping

AU - Zhang, Hu

AU - Xi, Hongsheng

PY - 2004

Y1 - 2004

N2 - We discuss the problems of discounted-eost performance optimization for a class of countable Markov control processes (CMCP) with (possibly) unbounded transition rates. The discounted Poisson equation is proposed for a CMCP, from which α -potential is defined. The optimality equation based on the α -potential is derived under the conditions weaker than those previously known.

AB - We discuss the problems of discounted-eost performance optimization for a class of countable Markov control processes (CMCP) with (possibly) unbounded transition rates. The discounted Poisson equation is proposed for a CMCP, from which α -potential is defined. The optimality equation based on the α -potential is derived under the conditions weaker than those previously known.

KW - Countable Markov control processes

KW - Discounted Poisson equations

KW - Discounted-cost performance criteria

KW - Unbounded transition rates

KW - α -potential

UR - https://www.scopus.com/pages/publications/4444318672

M3 - 论文

AN - SCOPUS:4444318672

SP - 357

EP - 359

T2 - WCICA 2004 - Fifth World Congress on Intelligent Control and Automation, Conference Proceedings

Y2 - 15 June 2004 through 19 June 2004

ER -

Markov decision problems with unbounded transition rates under discounted-cost performance criteria

Abstract

Conference

Keywords

Other files and links

Fingerprint

Cite this