25323 articles – 19562 references  [version française]
See detailed view Peer-reviewed conferences/proceedings
Genetic and Evolutionary Computation Conference, Montréal : Canada (2009)
Attached file list to this document: 
PDF
NOLH-GA.pdf(3.8 MB)
Optimizing Low-Discrepancy Sequences with an Evolutionary Algorithm
François-Michel De Rainville1, Christian Gagné1, Olivier Teytaud2, 3, 4, Denis Laurendeau1

Many elds rely on some stochastic sampling of a given com- plex space. Low-discrepancy sequences are methods aim- ing at producing samples with better space-lling properties than uniformly distributed random numbers, hence allow- ing a more ecient sampling of that space. State-of-the-art methods like nearly orthogonal Latin hypercubes and scram- bled Halton sequences are congured by permutations of in- ternal parameters, where permutations are commonly done randomly. This paper proposes the use of evolutionary al- gorithms to evolve these permutations, in order to optimize a discrepancy measure. Results show that an evolution- ary method is able to generate low-discrepancy sequences of signicantly better space-lling properties compared to sequences congured with purely random permutations.
1:  GEL-GIF - Département de génie électrique et de génie informatique
2:  INRIA Futurs - TAO
3:  LRI - Laboratoire de Recherche en Informatique
4:  INRIA Saclay - Ile de France - TAO