Continuous lunches are free!

Anne Auger, Olivier Teytaud

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

Abstract

This paper investigates extensions of No Free Lunch (NFL) theorems to countably infinite and uncountable infinite domains. The original NFLdue to Wolpert and Macready states that all search heuristics have the same performance when averaged over the uniform distribution over all possible functions. For infinite domains the extension of the concept of distribution over all possible functions involves measurability issues and stochastic process theory. For countably infinite domains, we prove that the natural extension of NFL theorems does not hold, but that a weaker form of NFL does hold, by stating the existence of non-trivial distributions of fitness leading to equal performance forall search heuristics. Our main result is that for continuous domains, NFL does not hold.

Original languageEnglish
Title of host publicationProceedings of GECCO 2007
Subtitle of host publicationGenetic and Evolutionary Computation Conference
Pages916-922
Number of pages7
DOIs
Publication statusPublished - 27 Aug 2007
Externally publishedYes
Event9th Annual Genetic and Evolutionary Computation Conference, GECCO 2007 - London, United Kingdom
Duration: 7 Jul 200711 Jul 2007

Publication series

NameProceedings of GECCO 2007: Genetic and Evolutionary Computation Conference

Conference

Conference9th Annual Genetic and Evolutionary Computation Conference, GECCO 2007
Country/TerritoryUnited Kingdom
CityLondon
Period7/07/0711/07/07

Keywords

  • Free-lunch
  • Kolmogorov's extension theorem
  • No-free-lunch

Fingerprint

Dive into the research topics of 'Continuous lunches are free!'. Together they form a unique fingerprint.

Cite this