Continuous lunches are free! - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2007

Continuous lunches are free!

Anne Auger
  • Function : Author
  • PersonId : 751513
  • IdHAL : anne-auger
Olivier Teytaud

Abstract

This paper investigates extensions of No Free Lunch (NFL) theorems to countably infinite and uncountable infinite do- mains. The original NFL due 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 measur- ability 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 distribution of fitness leading to equal performance for all search heuristics. Our main result is that for continuous domains, NFL does not hold.
Fichier principal
Vignette du fichier
cfl.pdf (134.75 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

inria-00173209 , version 1 (19-09-2007)

Identifiers

  • HAL Id : inria-00173209 , version 1

Cite

Anne Auger, Olivier Teytaud. Continuous lunches are free!. GECCO, 2007, London, United Kingdom. ⟨inria-00173209⟩
165 View
3548 Download

Share

Gmail Facebook X LinkedIn More