FILTRANE, a Fortran 95 filter-trust-region package for solving nonlinear least-squares and nonlinear feasibility problems

Nick Gould, Philippe Toint

    Research output: Contribution to journalArticlepeer-review

    101 Downloads (Pure)

    Abstract

    FILTRANE, a new Fortran 95 package for finding vectors satisfying general sets of nonlinear equations and/or inequalities, is presented. Several algorithmic variants are discussed and extensively compared on a set of CUTEr test problems, indicating that the default variant is both reliable and efficient. This discussion provides a first experimental study of the parameters inherent in filter algorithms. © 2007 ACM.
    Original languageEnglish
    Pages (from-to)3-25
    Number of pages23
    JournalACM Transactions on Mathematical Software
    Volume33
    Issue number1
    DOIs
    Publication statusPublished - 1 Mar 2007

    Keywords

    • numerical algorithms
    • nonlinear equations
    • nonlinear least-squares
    • nonlinear feasibility

    Fingerprint

    Dive into the research topics of 'FILTRANE, a Fortran 95 filter-trust-region package for solving nonlinear least-squares and nonlinear feasibility problems'. Together they form a unique fingerprint.

    Cite this