Abstract:
Evolutionary Markov chain Monte Carlo (EMCMC) is a class of algorithms obtained by merging Markov chain Monte Carlo algorithms with evolutionary computation methods. EMCM...Show MoreMetadata
Abstract:
Evolutionary Markov chain Monte Carlo (EMCMC) is a class of algorithms obtained by merging Markov chain Monte Carlo algorithms with evolutionary computation methods. EMCMC integrates techniques from the EC framework (population, recombination and selection) into the MCMC framework to increase the performance of the standard MCMC algorithms. In this paper, we show how to use recombination operators in EMCMC and how to combine them with other existing MCMC techniques (e.g. mutation and selection). We illustrate these principles by means of an example.
Published in: 2005 IEEE Congress on Evolutionary Computation
Date of Conference: 02-05 September 2005
Date Added to IEEE Xplore: 12 December 2005
Print ISBN:0-7803-9363-5