Free Search and Particle Swarm Optimisation Applied to Global Optimisation Numerical Tests from Two to Hundred Dimensions

Vesela Vasileva, Kalin Penev

    Research output: Chapter in Book/Report/Published conference proceedingChapterpeer-review

    Abstract

    This article presents an investigation on two real-value methods such as Free Search (FS) and Particle Swarm Optimisation (PSO) applied to global optimisation numerical tests. The objective is to identify how to facilitate assessment of heuristic, evolutionary, adaptive and other optimisation and search algorithms. Particular aim is to assess: (1) probability for success of given method; (2) abilities of given method for entire search space coverage; (3) dependence on initialisation; (4) abilities of given method to escape from trapping in local sub-optima; (5) abilities of explored methods to resolve multidimensional (one hundred dimensions) global optimisation tasks; (6) performance on two and hundred dimensional tasks; (7) minimal number of objective function calculation for resolving hundred dimensional tasks with acceptable level of precision. Achieved experimental results are presented and analysed. Discussion on FS and PSO essential characteristics concludes the article.
    Original languageEnglish
    Title of host publicationRecent Contributions in Intelligent Systems
    EditorsVassil Sgurev, Ronald Yager, Janusz Kacprzyk, Krassimir Atanassov
    Place of PublicationSwitzerland
    PublisherSpringer International Publishing AG
    Chapter17
    Pages313-337
    Number of pages24
    Volume657
    Edition1
    ISBN (Electronic)978-3-319-41438-6
    ISBN (Print)978-3-319-41437-9, 978-3-319-82354-6
    DOIs
    Publication statusPublished - 2017

    Publication series

    NameStudies in Computational Intelligence
    PublisherSpringer Nature Switzerland AG
    ISSN (Print)1860-949X

    Fingerprint Dive into the research topics of 'Free Search and Particle Swarm Optimisation Applied to Global Optimisation Numerical Tests from Two to Hundred Dimensions'. Together they form a unique fingerprint.

    Cite this