Lineability of the set of supporting vectors (Q2222834)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Lineability of the set of supporting vectors
scientific article

    Statements

    Lineability of the set of supporting vectors (English)
    0 references
    27 January 2021
    0 references
    Let \(T\) be a continuous linear operator between normed spaces \(X\) and \(Y\) over \(\mathbb{R}\). A vector \(x\in X\) is called a supporting vector of \(T\) if \(\left\Vert Tx\right\Vert =\left\Vert T\right\Vert \left\Vert x\right\Vert \); the set of supporting vectors of \(T\) is denoted by \textrm{suppv}\((T)\). A subset \(\mathcal{F}\) of \(X\) is called \(\lambda\)-lineable (= \(\lambda\)-spaceable) if \(\mathcal{F}\cup\{0\}\) contains a closed vector space of dimension \(\lambda\). Similarly \(\mathcal{F}\) is said to be \(\gamma\)-coneable if there is a cone \(C\subseteq\mathcal{F}\) containing \(\gamma\) linearly independent vectors (see for example the survey [\textit{L. Bernal-González} et al., Bull. Am. Math. Soc., New Ser. 51, No. 1, 71--130 (2014; Zbl 1292.46004)]). The present paper explores criteria for \(\lambda\)-lineability and \(\lambda\)-coneability of \textrm{suppv}\((T)\). The following are some examples. If \(x^{\ast}\) is a nonzero functional on \(X\) then \text{suppv}\((x^{\ast})\) is \(1\)- but not \(2\)-lineable and if \(X\) is strictly convex then it is also true that \text{suppv}\((x^{\ast})\) is \(1\)- but not \(2\)-coneable (strictly convex means that \(\left\Vert x+y\right\Vert =\left\Vert x\right\Vert +\left\Vert y\right\Vert \) only if \(x\) and \(y\) are linearly dependent). More generally for any nonzero \(T\) we have \text{suppv}\((T)\cap\ker(T)=\{0\}\); hence if \(\ker(T)\ \)has codimension \(n\) then \text{suppv}\((T)\) is at most \(n\)-lineable. If \(X\) and \(Y\) are Hilbert spaces then, under quite weak conditions, \text{suppv}\((T)\) is a subspace of \(X\). If \(X\) is a separable Hilbert space with orthonormal basis \((x_{n})_{n\in\mathbb{N}}\) and \((\rho_{n})\) is a bounded sequence of scalars with \(\rho:=\sup_{n}\rho_{n}\); then \(T(x):=\sum_{n}\rho_{n}(x|x_{n})x_{n}\) (\(x\in X\)) defines an operator \(T\) on \(X\) such that \text{suppv}\((T)\) is the closure of \(\left\{ x_{n}~|~\rho_{n}=\rho\right\} \). There are similar results on coneability and many examples to illustrate when (and when not) some of these results can be extended to spaces over \(\mathbb{C}\).
    0 references
    0 references
    lineability
    0 references
    coneability
    0 references
    supporting vector
    0 references
    norm-attaining functional
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references

    Identifiers