Convergence of simulated annealing using Foster-Lyapunov criteria

被引：18

作者：

Andrieu, C

Breyer, LA

Doucet, A

机构：

[1] Univ Bristol, Stat Grp, Dept Math, Bristol BS8 1TW, Avon, England

[2] Univ Lancaster, Fylde Coll, Dept Math & Stat, Lancaster LA1 4YF, England

[3] Univ Melbourne, Dept Elect & Elect Engn, Parkville, Vic 3010, Australia

来源：

JOURNAL OF APPLIED PROBABILITY | 2001年 / 38卷 / 04期

关键词：

simulated annealing; Markov chain Monte Carlo; optimization; Foster-Lyapunov; nonhomogeneous Markov chain;

D O I：

10.1017/S0021900200019173

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Simulated annealing is a popular and much studied method for maximizing functions on finite or compact spaces. For noncompact state spaces, the method is still sound, but convergence results are scarce, We show here how to prove convergence in such cases, for Markov chains satisfying suitable drift and minorization conditions.

引用

页码：975 / 994

页数：20

共 23 条

[1] Simulated annealing for maximum A Posteriori parameter estimation of hidden Markov models [J].