Evolutionary Markov chain Monte Carlo

M.M. Drugan; D. Thierens

doi:10.1007/978-3-540-24621-3_6

Evolutionary Markov chain Monte Carlo

M.M. Drugan, D. Thierens

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic › peer-review

14 Citations (SciVal)

Abstract

Markov chain Monte Carlo (MCMC) is a popular class of algorithms to sample from a complex distribution. A key issue in the design of MCMC algorithms is to improve the proposal mechanism and the mixing behaviour. This has led some authors to propose the use of a population of MCMC chains, while others go even further by integrating techniques from evolutionary computation (EC) into the MCMC framework. This merging of MCMC and EC leads to a class of algorithms, we call Evolutionary Markov Chain Monte Carlo (EMCMC). In this paper we first survey existing EMCMC algorithms and categorise them in two classes: family-competitive EMCMC and population-driven EMCMC. Next, we introduce, the Elitist Coupled Acceptance rule and the Fitness Ordered Tempering algorithm.

Original language	English
Title of host publication	Artificial Evolution
Subtitle of host publication	6th International Conference, Evolution Artificielle, EA 2003, Marseilles, France, October 27-30, 2003, Revised Selected Papers
Editors	P. Liardet, P. Collet, C. Fonlupt, E. Lutton, M. Schoenauer
Place of Publication	Berlin
Publisher	Springer
Pages	63-76
Number of pages	14
ISBN (Print)	978-3-540-21523-3
DOIs	https://doi.org/10.1007/978-3-540-24621-3_6
Publication status	Published - 2004
Externally published	Yes

Publication series

Name	LNCS
Volume	2936

Keywords

algorithms

Access to Document

10.1007/978-3-540-24621-3_6

http://www.mendeley.com/research/evolutionary-markov-chain-monte-carlo

Cite this

Drugan, M. M., & Thierens, D. (2004). Evolutionary Markov chain Monte Carlo. In P. Liardet, P. Collet, C. Fonlupt, E. Lutton, & M. Schoenauer (Eds.), Artificial Evolution: 6th International Conference, Evolution Artificielle, EA 2003, Marseilles, France, October 27-30, 2003, Revised Selected Papers (pp. 63-76). (LNCS; Vol. 2936). Springer. https://doi.org/10.1007/978-3-540-24621-3_6

@inbook{4cc89db64e054cd692aa9960774a484a,

title = "Evolutionary Markov chain Monte Carlo",

abstract = "Markov chain Monte Carlo (MCMC) is a popular class of algorithms to sample from a complex distribution. A key issue in the design of MCMC algorithms is to improve the proposal mechanism and the mixing behaviour. This has led some authors to propose the use of a population of MCMC chains, while others go even further by integrating techniques from evolutionary computation (EC) into the MCMC framework. This merging of MCMC and EC leads to a class of algorithms, we call Evolutionary Markov Chain Monte Carlo (EMCMC). In this paper we first survey existing EMCMC algorithms and categorise them in two classes: family-competitive EMCMC and population-driven EMCMC. Next, we introduce, the Elitist Coupled Acceptance rule and the Fitness Ordered Tempering algorithm.",

keywords = "algorithms",

author = "M.M. Drugan and D. Thierens",

year = "2004",

doi = "10.1007/978-3-540-24621-3_6",

language = "English",

isbn = "978-3-540-21523-3",

series = "LNCS",

publisher = "Springer",

pages = "63--76",

editor = "P. Liardet and P. Collet and C. Fonlupt and E. Lutton and { Schoenauer}, M.",

booktitle = "Artificial Evolution",

address = "Germany",

}

Drugan, MM & Thierens, D 2004, Evolutionary Markov chain Monte Carlo. in P Liardet, P Collet, C Fonlupt, E Lutton & M Schoenauer (eds), Artificial Evolution: 6th International Conference, Evolution Artificielle, EA 2003, Marseilles, France, October 27-30, 2003, Revised Selected Papers. LNCS, vol. 2936, Springer, Berlin, pp. 63-76. https://doi.org/10.1007/978-3-540-24621-3_6

Evolutionary Markov chain Monte Carlo. / Drugan, M.M.; Thierens, D.
Artificial Evolution: 6th International Conference, Evolution Artificielle, EA 2003, Marseilles, France, October 27-30, 2003, Revised Selected Papers. ed. / P. Liardet; P. Collet; C. Fonlupt; E. Lutton; M. Schoenauer. Berlin: Springer, 2004. p. 63-76 (LNCS; Vol. 2936).

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic › peer-review

TY - CHAP

T1 - Evolutionary Markov chain Monte Carlo

AU - Drugan, M.M.

AU - Thierens, D.

PY - 2004

Y1 - 2004

N2 - Markov chain Monte Carlo (MCMC) is a popular class of algorithms to sample from a complex distribution. A key issue in the design of MCMC algorithms is to improve the proposal mechanism and the mixing behaviour. This has led some authors to propose the use of a population of MCMC chains, while others go even further by integrating techniques from evolutionary computation (EC) into the MCMC framework. This merging of MCMC and EC leads to a class of algorithms, we call Evolutionary Markov Chain Monte Carlo (EMCMC). In this paper we first survey existing EMCMC algorithms and categorise them in two classes: family-competitive EMCMC and population-driven EMCMC. Next, we introduce, the Elitist Coupled Acceptance rule and the Fitness Ordered Tempering algorithm.

AB - Markov chain Monte Carlo (MCMC) is a popular class of algorithms to sample from a complex distribution. A key issue in the design of MCMC algorithms is to improve the proposal mechanism and the mixing behaviour. This has led some authors to propose the use of a population of MCMC chains, while others go even further by integrating techniques from evolutionary computation (EC) into the MCMC framework. This merging of MCMC and EC leads to a class of algorithms, we call Evolutionary Markov Chain Monte Carlo (EMCMC). In this paper we first survey existing EMCMC algorithms and categorise them in two classes: family-competitive EMCMC and population-driven EMCMC. Next, we introduce, the Elitist Coupled Acceptance rule and the Fitness Ordered Tempering algorithm.

KW - algorithms

U2 - 10.1007/978-3-540-24621-3_6

DO - 10.1007/978-3-540-24621-3_6

M3 - Chapter

SN - 978-3-540-21523-3

T3 - LNCS

SP - 63

EP - 76

BT - Artificial Evolution

A2 - Liardet, P.

A2 - Collet, P.

A2 - Fonlupt, C.

A2 - Lutton, E.

A2 - Schoenauer, M.

PB - Springer

CY - Berlin

ER -

Evolutionary Markov chain Monte Carlo

Abstract

Publication series

Keywords

Access to Document

Fingerprint

Cite this