On the convergence of localisation search

David Bulger, Graham Wood

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Localisation Search, a generalisation of Pure Adaptive Search, is a stochastic global optimisation algorithm in which iterates are chosen randomly from a superset, or localisation, of the improving region. Recent theoretical results are applied to determine how closely the localisation must track the improving region in order that the number of function evaluations to convergence increases at most polynomially with dimension, for Lipschitz objective functions and uniformly bounded domains.
Original languageEnglish
Title of host publicationState of the Art in Global Optimisation
Subtitle of host publicationComputational Methods and Applications
EditorsChris Floudas, Panos Pardalos
Place of PublicationBoston
PublisherSpringer, Springer Nature
Pages227-233
Number of pages7
Volume7
ISBN (Electronic)9781461334378
ISBN (Print)9781461334392
DOIs
Publication statusPublished - 1996
Externally publishedYes

Publication series

NameNonconvex Optimization and Its Applications
PublisherSpringer
ISSN (Print)1571-568X

Keywords

  • global optimization
  • algorithm complexity
  • random search
  • adaptive search
  • localisation
  • Lipschitz function

Fingerprint Dive into the research topics of 'On the convergence of localisation search'. Together they form a unique fingerprint.

  • Cite this

    Bulger, D., & Wood, G. (1996). On the convergence of localisation search. In C. Floudas, & P. Pardalos (Eds.), State of the Art in Global Optimisation: Computational Methods and Applications (Vol. 7, pp. 227-233). (Nonconvex Optimization and Its Applications). Boston: Springer, Springer Nature. https://doi.org/10.1007/978-1-4613-3437-8_15