La lecture en ligne est gratuite
Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres
Télécharger Lire

[inria-00377093, v1] Application of a simple binary genetic algorithm to a noiseless testbed benchmark

6 pages
Author manuscript, published in "Genetic and Evolutionary Computation Conference (GECCO) (2009)"Application of a simple binary genetic algorithm to anoiseless testbed benchmarkMiguel NicolauINRIA Saclay - Île-de-FranceLRI - Université Paris SudParis, Francemiguel.nicolau@lri.frABSTRACT this problem, a new breed of genetic algorithms, the so-called Messy-GAs [5] have been developed. These GAs areOne of the earliest evolutionary computation algorithms,certainly better suited for higher order problems; however,the genetic algorithm, is applied to the noise-free BBOBthey are quite complex and hard to deploy. As a result,2009 testbed. It is adapted to the continuous domain bythe simplicity of developing and applying a simple GA to aincreasing the number of bits encoding each variable, un-variety of problems remains one of its biggest strengths.til a desired resolution is possible to achieve. Good resultsThecurrentpaperthereforeadaptstheoriginal, simplebi-and scaling are obtained for separable functions, but poornaryGAtoacontinuousdomainproblem,andreportsonitsperformance is achieved on the other functions, particularlyperformance. Although not achieving stellar performance,ill-conditioned functions. Overall running times remain fastparticularly incomparison with more recentandperformingthroughout.algorithms,theresultsobtainedareneverthelessremarkable,particularly in separable functions.Categories and Subject DescriptorsG.1.6 [Numerical Analysis]: ...
Voir plus Voir moins
 ˘ˇ ˆ˙˝  ˛˚ˇ ˝ ˜ ˙ !˘ !"ˇ #