PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Optimizing Low-Discrepancy Sequences with an Evolutionary Algorithm
Francois-Michel De Rainville, Christian Gagné, Olivier Teytaud and Denis Laurendeau
In: Gecco 2009(2009).

Abstract

Many fields rely on some stochastic sampling of a given com- plex space. Low-discrepancy sequences are methods aim- ing at producing samples with better space-filling properties than uniformly distributed random numbers, hence allow- ing a more efficient sampling of that space. State-of-the-art methods like nearly orthogonal Latin hypercubes and scram- bled Halton sequences are configured 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 significantly better space-filling properties compared to sequences configured with purely random permutations.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Talk)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Learning/Statistics & Optimisation
ID Code:6888
Deposited By:Olivier Teytaud
Deposited On:09 April 2010