Best constant approximants in Lorentz spaces (Q707204)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Best constant approximants in Lorentz spaces
scientific article

    Statements

    Best constant approximants in Lorentz spaces (English)
    0 references
    9 February 2005
    0 references
    Let \(M_{0}\) be the class of all real extended \(\mu \)-measurable functions on \([0,1]\), where \(\mu \) is the Lebesgue measure. For \(f\in M_{0}\), denote by \( \mu _{f}(\lambda )\) the distribution function of \(f\), that is, \(\mu _{f}(\lambda )=\mu (\{x\in [ 0,1]:\mid f(x)\mid >\lambda \})\), (\( \lambda \geq 0\)), and by \(f^{\ast }(t)=\inf \{\lambda :\mu _{f}(\lambda )\) \( \leq t\}\), (\(t\geq 0)\) its decreasing rearrangement. Let \(\omega :(0,1]\rightarrow (0,\infty )\) be a weigth function, non-increasing and locally integrable with respect to \(\mu \). For \(1\leq q<\infty \), the authors consider the Lorentz space \(L^{\omega ,q}=\{f\in M_{0}:\parallel f\parallel _{\omega ,q}<\infty \}\), where \[ \parallel f\parallel _{\omega ,q}=\left( \int_{0}^{1}\omega (t)\left( f^{\ast }(t)\right) ^{q}d\mu (t)\right) ^{1/q}. \] For \(\;f\in L^{\omega ,q}\), let \(C_{f}:=\{c\in \mathbb{R:}\) \(\parallel f-c\parallel =\inf_{k\in \mathbb{R}}\parallel f-k\parallel \}\). Each element of \(C_{f}\) is called a best constant approximant of \(f\). It is known that \( C_{f}\) is a non-empty and compact set of \(\mathbb{R}\). Denote \(\underline{f} =\min (C_{f})\) and \(\overline{f}=\max (C_{f})\) and consider the best approximation operator defined by \(T(f)=C_{f}\). The authors give a characterization of the best constant approximants for a simple function, \(f\), and they stablish a way to obtain the best constant approximant maximum, \(\overline{f}\), and minimum, \(\underline{f}\). Also they study the monotony of the best constant approximation operator \(T\) on the set of the simple functions, in the sense of \textit{D. Landers} and \textit{L. Rogge} [Z. Wahrsch. Verw. Geb. 51, 29--389 (1982; Zbl 0422.60039)], by proving that if \(f,g\) are two simple functions, \(f\leq g\), \(c\in C_{f}\) and \(d\in C_{g}\), then \( c\vee d\in C_{g}\) and \(c\wedge g\in C_{f}\), where \(c\vee d=\max \{f,g\}\) and \(c\wedge g=\min \{f,g\}\).
    0 references
    0 references
    best constant approximants
    0 references
    Lorentz spaces
    0 references
    monotony
    0 references
    0 references
    0 references
    0 references