System Identification Using Evolutionary Markov Chain
Monte Carlo
B. Zhang, and D. Cho. Journal of Systems Architecture, 47 (7):
587--599(July 2001)
Abstract
System identification involves determination of the
functional structure of a target system that underlies
the observed data. In this paper, we present a
probabilistic evolutionary method that optimises system
architectures for the identification of unknown target
systems. The method is distinguished from existing
evolutionary algorithms in that the individuals are
generated from a probability distribution as in Markov
chain Monte Carlo (MCMC). It is also distinguished from
conventional MCMC methods in that the search is
population-based as in standard evolutionary
algorithms. The effectiveness of this hybrid of
evolutionary computation and MCMC is tested on a
practical problem, i.e., evolving neural net
architectures for the identification of nonlinear
dynamic systems. Experimental evidence supports that
evolutionary MCMC (or eMCMC) exploits the efficiency of
simple evolutionary algorithms while maintaining the
robustness of MCMC methods and outperforms either
approach used alone.
%0 Journal Article
%1 Zhang:2001:JSA
%A Zhang, Byoung-Tak
%A Cho, Dong-Yeon
%D 2001
%J Journal of Systems Architecture
%K Carlo Markov Monte System chain identification,
%N 7
%P 587--599
%T System Identification Using Evolutionary Markov Chain
Monte Carlo
%V 47
%X System identification involves determination of the
functional structure of a target system that underlies
the observed data. In this paper, we present a
probabilistic evolutionary method that optimises system
architectures for the identification of unknown target
systems. The method is distinguished from existing
evolutionary algorithms in that the individuals are
generated from a probability distribution as in Markov
chain Monte Carlo (MCMC). It is also distinguished from
conventional MCMC methods in that the search is
population-based as in standard evolutionary
algorithms. The effectiveness of this hybrid of
evolutionary computation and MCMC is tested on a
practical problem, i.e., evolving neural net
architectures for the identification of nonlinear
dynamic systems. Experimental evidence supports that
evolutionary MCMC (or eMCMC) exploits the efficiency of
simple evolutionary algorithms while maintaining the
robustness of MCMC methods and outperforms either
approach used alone.
@article{Zhang:2001:JSA,
abstract = {System identification involves determination of the
functional structure of a target system that underlies
the observed data. In this paper, we present a
probabilistic evolutionary method that optimises system
architectures for the identification of unknown target
systems. The method is distinguished from existing
evolutionary algorithms in that the individuals are
generated from a probability distribution as in Markov
chain Monte Carlo (MCMC). It is also distinguished from
conventional MCMC methods in that the search is
population-based as in standard evolutionary
algorithms. The effectiveness of this hybrid of
evolutionary computation and MCMC is tested on a
practical problem, i.e., evolving neural net
architectures for the identification of nonlinear
dynamic systems. Experimental evidence supports that
evolutionary MCMC (or eMCMC) exploits the efficiency of
simple evolutionary algorithms while maintaining the
robustness of MCMC methods and outperforms either
approach used alone.},
added-at = {2008-06-19T17:46:40.000+0200},
author = {Zhang, Byoung-Tak and Cho, Dong-Yeon},
biburl = {https://www.bibsonomy.org/bibtex/221261ef86b7523491a3e5ab6e04ccdee/brazovayeye},
interhash = {c8029b1c8c92b5555a1d8b950f88542a},
intrahash = {21261ef86b7523491a3e5ab6e04ccdee},
journal = {Journal of Systems Architecture},
keywords = {Carlo Markov Monte System chain identification,},
month = {July},
number = 7,
pages = {587--599},
timestamp = {2008-06-19T17:55:24.000+0200},
title = {System Identification Using Evolutionary Markov Chain
Monte Carlo},
volume = 47,
year = 2001
}