Sequential improvement for robust optimization using an uncertainty measure for radial basis functions

    Research output: Contribution to journalArticleAcademicpeer-review

    14 Citations (Scopus)
    67 Downloads (Pure)

    Abstract

    The performance of the sequential metamodel based optimization procedure depends strongly on the chosen building blocks for the algorithm, such as the used metamodeling method and sequential improvement criterion. In this study, the effect of these choices on the efficiency of the robust optimization procedure is investigated. A novel sequential improvement criterion for robust optimization is proposed, as well as an improved implementation of radial basis function interpolation suitable for sequential optimization. The leave-one-out cross-validation measure is used to estimate the uncertainty of the radial basis function metamodel. The metamodeling methods and sequential improvement criteria are compared, based on a test with Gaussian random fields as well as on the optimization of a strip bending process with five design variables and two noise variables. For this process, better results are obtained in the runs with the novel sequential improvement criterion as well as with the novel radial basis function implementation, compared to the runs with conventional sequential improvement criteria and kriging interpolation.
    Original languageEnglish
    Pages (from-to)1345-1363
    Number of pages19
    JournalStructural and multidisciplinary optimization
    Volume55
    Issue number4
    DOIs
    Publication statusPublished - Apr 2017

    Fingerprint

    Uncertainty Measure
    Robust Optimization
    Radial Functions
    Basis Functions
    Metamodeling
    Interpolation
    Metamodel
    Optimization
    Interpolate
    Gaussian Random Field
    Kriging
    Uncertainty
    Cross-validation
    Building Blocks
    Strip
    Estimate

    Keywords

    • IR-102280
    • METIS-319101

    Cite this

    @article{06063d85d789420285659a10ef2bfd47,
    title = "Sequential improvement for robust optimization using an uncertainty measure for radial basis functions",
    abstract = "The performance of the sequential metamodel based optimization procedure depends strongly on the chosen building blocks for the algorithm, such as the used metamodeling method and sequential improvement criterion. In this study, the effect of these choices on the efficiency of the robust optimization procedure is investigated. A novel sequential improvement criterion for robust optimization is proposed, as well as an improved implementation of radial basis function interpolation suitable for sequential optimization. The leave-one-out cross-validation measure is used to estimate the uncertainty of the radial basis function metamodel. The metamodeling methods and sequential improvement criteria are compared, based on a test with Gaussian random fields as well as on the optimization of a strip bending process with five design variables and two noise variables. For this process, better results are obtained in the runs with the novel sequential improvement criterion as well as with the novel radial basis function implementation, compared to the runs with conventional sequential improvement criteria and kriging interpolation.",
    keywords = "IR-102280, METIS-319101",
    author = "Havinga, {Gosse Tjipke} and {van den Boogaard}, {Antonius H.} and G. Klaseboer",
    year = "2017",
    month = "4",
    doi = "10.1007/s00158-016-1572-5",
    language = "English",
    volume = "55",
    pages = "1345--1363",
    journal = "Structural and multidisciplinary optimization",
    issn = "1615-147X",
    publisher = "Springer",
    number = "4",

    }

    Sequential improvement for robust optimization using an uncertainty measure for radial basis functions. / Havinga, Gosse Tjipke; van den Boogaard, Antonius H.; Klaseboer, G.

    In: Structural and multidisciplinary optimization, Vol. 55, No. 4, 04.2017, p. 1345-1363.

    Research output: Contribution to journalArticleAcademicpeer-review

    TY - JOUR

    T1 - Sequential improvement for robust optimization using an uncertainty measure for radial basis functions

    AU - Havinga, Gosse Tjipke

    AU - van den Boogaard, Antonius H.

    AU - Klaseboer, G.

    PY - 2017/4

    Y1 - 2017/4

    N2 - The performance of the sequential metamodel based optimization procedure depends strongly on the chosen building blocks for the algorithm, such as the used metamodeling method and sequential improvement criterion. In this study, the effect of these choices on the efficiency of the robust optimization procedure is investigated. A novel sequential improvement criterion for robust optimization is proposed, as well as an improved implementation of radial basis function interpolation suitable for sequential optimization. The leave-one-out cross-validation measure is used to estimate the uncertainty of the radial basis function metamodel. The metamodeling methods and sequential improvement criteria are compared, based on a test with Gaussian random fields as well as on the optimization of a strip bending process with five design variables and two noise variables. For this process, better results are obtained in the runs with the novel sequential improvement criterion as well as with the novel radial basis function implementation, compared to the runs with conventional sequential improvement criteria and kriging interpolation.

    AB - The performance of the sequential metamodel based optimization procedure depends strongly on the chosen building blocks for the algorithm, such as the used metamodeling method and sequential improvement criterion. In this study, the effect of these choices on the efficiency of the robust optimization procedure is investigated. A novel sequential improvement criterion for robust optimization is proposed, as well as an improved implementation of radial basis function interpolation suitable for sequential optimization. The leave-one-out cross-validation measure is used to estimate the uncertainty of the radial basis function metamodel. The metamodeling methods and sequential improvement criteria are compared, based on a test with Gaussian random fields as well as on the optimization of a strip bending process with five design variables and two noise variables. For this process, better results are obtained in the runs with the novel sequential improvement criterion as well as with the novel radial basis function implementation, compared to the runs with conventional sequential improvement criteria and kriging interpolation.

    KW - IR-102280

    KW - METIS-319101

    U2 - 10.1007/s00158-016-1572-5

    DO - 10.1007/s00158-016-1572-5

    M3 - Article

    VL - 55

    SP - 1345

    EP - 1363

    JO - Structural and multidisciplinary optimization

    JF - Structural and multidisciplinary optimization

    SN - 1615-147X

    IS - 4

    ER -