Two characterizations of smooth norms (Q1306493)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Two characterizations of smooth norms
scientific article

    Statements

    Two characterizations of smooth norms (English)
    0 references
    0 references
    0 references
    4 January 2001
    0 references
    Let \(\gamma\) be a norm with unit ball \(B = \{x\in\mathbb R^n: \gamma(x)\leqslant 1\}\), \(n\geqslant 2.\) The norm \(\gamma\) and its unit ball \(B\) are said to be smooth if for each \(t\in\mathbb R^n\) with \(\gamma(t) = 1\) there exists exactly one supporting hyperplane of \(B\) at \(t.\) The authors give the characterizations of smooth norms. The norm \(\gamma\) is smooth iff for all weighted point sets \(P\) not contained in a hyperplane each median hyperplane (with respect to \(\gamma\) and \(P\)) is the affine hull of \(n\) points from \(P,\) i.e., each median hyperplane satisfies the incidence criterion. The norm \(\gamma\) is smooth iff for all weighted point sets \(P\) not contained in a hyperplane each center hyperplane (with respect to \(\gamma\) and \(P\)) is at maximum distance to \(n + 1\) affinely independent points from \(P,\) i.e., every center hyperplane is blocked by \(P.\)
    0 references
    0 references
    0 references
    0 references
    0 references
    center hyperplane
    0 references
    centered hyperplane
    0 references
    distance
    0 references
    \(k\)-set problem
    0 references
    median hyperplane
    0 references
    Minkowski space
    0 references
    minsum hyperplane
    0 references
    norm
    0 references
    numerical approximation
    0 references
    point set width problem
    0 references
    robust statistics
    0 references
    smooth norm
    0 references