Benchmarking the (1+1)-ES with One-Fifth Success rule on the BBOB-2009 Noisy Testbed

Anne Auger 1, 2
1 TAO - Machine Learning and Optimisation
LRI - Laboratoire de Recherche en Informatique, UP11 - Université Paris-Sud - Paris 11, Inria Saclay - Ile de France, CNRS - Centre National de la Recherche Scientifique : UMR8623
Abstract : The (1+1)-ES with one-fifth success rule is one of the first and simplest stochastic algorithm proposed for optimization on a continuous search space in a black-box scenario. In this paper, we benchmark an independent-restart (1+1)-ES with one-fifth success rule on the BBOB-2009 noisy testbed. The maximum number of function evaluations used equals $10^{6}$ times the dimension of the search space. The algorithm could only solve 3 functions with moderate noise in 5-D and 2 functions in 20-D.
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/inria-00430516
Contributor : Anne Auger <>
Submitted on : Sunday, November 8, 2009 - 1:54:23 PM
Last modification on : Thursday, April 5, 2018 - 12:30:12 PM
Long-term archiving on : Thursday, June 17, 2010 - 7:45:03 PM

File

Auger-wk2038.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00430516, version 1

Collections

Citation

Anne Auger. Benchmarking the (1+1)-ES with One-Fifth Success rule on the BBOB-2009 Noisy Testbed. ACM-GECCO Genetic and Evolutionary Computation Conference, Jul 2009, Montreal, Canada. ⟨inria-00430516⟩

Share

Metrics

Record views

265

Files downloads

213