Research on the Bayesian optimization algorithms (IlliGAL report No 200010

fpelikandeggilligalgeuiucedu This paper summarizes our recent research on the Bayesian optimization algorithm BOA and outlines the directions our research in this area has been following It settles the algorithm in the problem decomposition framework used often to understand the complex behavior of genetic algorithms It provides the most important research issues to tackle and reviews our recent progress in each of these areas The use of search methods that build models of promising solutions found so far and use the built models to guide the further search has recently shown to result in powerful methods that resolve many of the problems of evolutionary methods identied by the genetic and evolutionary computation community over the past decade