A dynamic programming approach for finite Markov processes and algorithms for the calculation of the limit matrix in Markov chains

被引：1

作者：

Pickl, Stefan ^{[1
]}

Lozovanu, Dmitrii ^{[2
]}

机构：

[1] Univ Bundeswehr, Inst Theoret Informat Math & Operat Res, Fak Informat, Munich, Germany

[2] Acad Sci, Inst Math & Comp Sci, MD-2028 Kishinev, Moldova

来源：

OPTIMIZATION | 2011年 / 60卷 / 10-11期

关键词：

discrete Markov process; probability of state transition; limit state matrix; dynamic programming; polynomial time algorithm;

D O I：

10.1080/02331934.2011.627333

中图分类号：

C93 [管理学]; O22 [运筹学];

学科分类号：

070105 ; 12 ; 1201 ; 1202 ; 120202 ;

摘要：

New calculation procedures for finding the probabilities of state transitions of the system in discrete Markov processes based on dynamic programming are developed and polynomial time algorithms for determining the limit state matrix in such processes are proposed. Computational complexity aspects and possible applications of the proposed algorithms for the stochastic optimization problems are characterized.

引用

页码：1339 / 1358

页数：20