A sparse quasi newton update derived variationally with a nondiagonally weighted frobenius norm

    Research output: Contribution to journalArticlepeer-review

    1 Downloads (Pure)

    Abstract

    The problem of symmetric sparse updating is considered from a variational point of view and a new class of sparse symmetric quasi Newton updating formulae is derived. This class results from the use of a nondiagonally weighted Frobenius norm. The computation of the update involves only one positive definite and symmetric linear system that has the same sparsity pattern as the problem itself.

    Original languageEnglish
    Pages (from-to)425-433
    Number of pages9
    JournalMathematics Of Computation
    Volume37
    Issue number156
    DOIs
    Publication statusPublished - 1 Jan 1981

    Keywords

    • Matrix updating
    • Nonlinear optimization
    • Sparsity

    Fingerprint Dive into the research topics of 'A sparse quasi newton update derived variationally with a nondiagonally weighted frobenius norm'. Together they form a unique fingerprint.

    Cite this