On the swapping algorithm

被引：23

作者：

Madras, N ^{[1
]}

Zheng, ZR ^{[1
]}

机构：

[1] York Univ, Dept Math & Stat, N York, ON M3J 1P3, Canada

来源：

RANDOM STRUCTURES & ALGORITHMS | 2003年 / 22卷 / 01期

关键词：

Markov chain Monte Carlo; spectral gap; metropolis-coupled Markov chains; meanfield Ising model; decomposition;

D O I：

10.1002/rsa.10066

中图分类号：

TP31 [计算机软件];

学科分类号：

081202 ; 0835 ;

摘要：

The Metropolis-coupled Markov chain method (or "Swapping Algorithm") is an empirically successful hybrid Monte Carlo algorithm. It alternates between standard transitions on parallel versions of the system at different parameter values, and swapping two versions. We prove rapid mixing for two bimodal examples, including the mean-field Ising model. (C) 2002 Wiley Periodicals, Inc.

引用

页码：66 / 97

页数：32

共 30 条

[1]

[Anonymous], THESIS YORK U

[2]

Binder K., 1992, MONTE CARLO SIMULATI

[3]

CARACCIOLO S, 1992, UNPUB 2 REMARKS SIMU

[4] Estimation of spectral gap for elliptic operators [J].

Chen, MF ;

Wang, FY .

TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 1997, 349 (03) :1239-1267

[5]

Chen Mufa, 1996, Acta Mathematica Sinica. New Series, V12, P337

[6] Mixing properties of the Swendsen-Wang process on the complete graph and narrow grids [J].

Cooper, C ;

Dyer, ME ;

Frieze, AM ;

Rue, R .

JOURNAL OF MATHEMATICAL PHYSICS, 2000, 41 (03) :1499-1527

[7]

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

[8] What do we know about the metropolis algorithm? [J].

Diaconis, P ;

Saloff-Coste, L .

JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 1998, 57 (01) :20-36

[9]

Diaconis P., 1991, Ann. Appl. Probab., P36

[10]

Diaconis P., 1993, Ann. Appl. Probab., V3, P696, DOI [10.1214/aoap/1177005359, DOI 10.1214/AOAP/1177005359]

← 1 2 3 →