What do we know about the metropolis algorithm?

被引：88

作者：

Diaconis, P ^{[1
]}

Saloff-Coste, L

机构：

[1] Stanford Univ, Dept Stat, Stanford, CA 94305 USA

[2] Univ Toulouse 3, CNRS, F-31062 Toulouse, France

[3] Cornell Univ, Dept Math, Ithaca, NY 14853 USA

来源：

JOURNAL OF COMPUTER AND SYSTEM SCIENCES | 1998年 / 57卷 / 01期

基金：

美国国家科学基金会;

关键词：

D O I：

10.1006/jcss.1998.1576

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

The Metropolis algorithm is a widely used procedure for sampling from a specified distribution on a large finite set. We survey what is rigorously known about running times. This includes work from statistical physics, computer science, probability, and statistics. Some new results (Propositions 6.1-6.5) are given as an illustration of the geometric theory of Markov chains. (C) 1998Academic Press.

引用

页码：20 / 36

页数：17

共 47 条

[1]

[Anonymous], CONT MATH

[2]

Barletta N., 1995, PROBABILITY PHASE TR, P265

[3]

Belsley E.D., 1993, Ph.D. thesis

[4]

Chung F, 1997, C BOARD MATH SCI AM

[5]

Chung F. R. K., 1995, COMB PROBAB COMPUT, V4, P11, DOI DOI 10.1017/S0963548300001449

[6]

CRITCHLOW D, 1985, SPRINGER LECT NOTE B, V4

[7] L-2 CONVERGENCE OF TIME NONHOMOGENEOUS MARKOV PROCESSES: I. SPECTRAL ESTIMATES [J].

Deuschel, Jean-Dominique ;

Mazza, Christian .

ANNALS OF APPLIED PROBABILITY, 1994, 4 (04) :1012-1056

[8]

Diaconis P, 1996, ANN APPL PROBAB, V6, P695

[9] Nash inequalities for finite Markov chains [J].

Diaconis, P ;

SaloffCoste, L .

JOURNAL OF THEORETICAL PROBABILITY, 1996, 9 (02) :459-510

[10]

Diaconis P., 1988, GROUP REPRESENTATION

← 1 2 3 4 5 →