Black-box complexity: Breaking the O(n log n) barrier of leadingones

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We show that the unrestricted black-box complexity of the n-dimensional XOR- and permutation-invariant LeadingOnes function class is O(n log(n) / loglogn). This shows that the recent natural looking O(nlogn) bound is not tight. The black-box optimization algorithm leading to this bound can be implemented in a way that only 3-ary unbiased variation operators are used. Hence our bound is also valid for the unbiased black-box complexity recently introduced by Lehre and Witt. The bound also remains valid if we impose the additional restriction that the black-box algorithm does not have access to the objective values but only to their relative order (ranking-based black-box complexity).

Original languageEnglish
Title of host publicationArtificial Evolution - 10th International Conference Evolution Artificielle, EA 2011, Revised Selected Papers
Pages205-216
Number of pages12
DOIs
Publication statusPublished - 14 Dec 2012
Externally publishedYes
Event10th International Conference on Evolution Artificielle, EA 2011 - Angers, France
Duration: 24 Oct 201126 Oct 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7401 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Evolution Artificielle, EA 2011
Country/TerritoryFrance
CityAngers
Period24/10/1126/10/11

Keywords

  • Algorithms
  • black-box complexity
  • query complexity
  • runtime analysis
  • theory

Fingerprint

Dive into the research topics of 'Black-box complexity: Breaking the O(n log n) barrier of leadingones'. Together they form a unique fingerprint.

Cite this