The Riesz potential operator in optimal couples of rearrangement invariant spaces (Q537682)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The Riesz potential operator in optimal couples of rearrangement invariant spaces
scientific article

    Statements

    The Riesz potential operator in optimal couples of rearrangement invariant spaces (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    20 May 2011
    0 references
    The Euclidean \(n\)-dimensional space is denoted by \(\mathbb{R}^n\), the space of locally Lebesgue integrable functions is denoted by \(L_{\text{loc}}\), the space of nonnegative functions in \(L_{\text{loc}}\) is denoted by \(L\), and the Riesz potential operator \(P_\alpha\) is defined for \(0< \alpha< n\) by \[ P_\alpha(f)(x)= \int_{\mathbb{R}^n} f(y)|x- y|^{\alpha- n}dy, \] where for \(x= (x_1,x_2,\dots, x_n)\in\mathbb{R}^n\),\ \(|x|= (x^2_1+ x^2_2+\cdots+ x^2_n)^{1/2}\). If \(m\) denotes Lebesgue measure on \(\mathbb{R}^n\), then the distribution function \(m_f(t)\), \(t> 0\), is identified as \(m(\{x\in\mathbb{R}^n: |f(x)|> t\})\), and the nonincreasing rearrangement \(f^*\) is defined by \(f^*(s)= \text{inf}\{t> 0: m_f(t)\leq s\}\). If \(E\) is a `rearrangement invariant' Banach function space in \(L^1+ L^\infty\), then a quasi-norm \(\|.\|_E\) is defined on \(E\) in terms of a quasi-norm \(\rho_E\) by \(\| f\|_E= \rho_E(f^*)\). If \(g_u\) is defined by \(g_u(t)= g_u(t/u)\), \(h_E(u)= \sup\{\rho_E(g^*_u)/\rho_E(g^*), g\in L\}\), then the Boyd indices \(\alpha_E\), \(\beta_E\), are defined by \[ \alpha_E= \sup\{(\log h_E(t)/\log t), 0< t< 1\},\;\beta_E= \text{inf}\{(\log h_E(t)/\log t), 1< t< \infty\}. \] The quasi-norm \(\rho_E\) is said to be monotone if \(\int^t_0 g^*_1(s)\,ds\leq \int^t_0 g^*_2(s)\,ds\) implies that \(\rho_E(g^*_1)\leq \rho_E(g^*_2)\). If \(\rho_E\) is monotone, rearrangement invariant with \(\alpha_E> \alpha/n\), \(\beta_E< 1\) and \(\rho_G\) is monotone and \(\beta_G< 1-(\alpha/n)\), then \(\rho_E\), \(\rho_G\) are said to be admissible if \(\rho_G(P_\alpha(f)^{**})/\rho_E(f^*)\leq \kappa\) for some constant \(\kappa\). The optimal target quasi-norm, \(\rho_G(E)\), satisfies \(\rho_G(g^*)/\rho_{G(E)}(g^*)\leq \kappa\), and the optimal domain quasi-norm \(\rho_{E(G)}(g^*)/\rho_E(g^*)\leq \kappa\), \(g\in L\). In the main results of this paper, conditions are provided under which quasi-norms are admissible, target quasi-norms are optimal and domain quasi-norms are optimal. In particular, it is shown that the quasi-norms \(\rho_E\), \(\rho_G\) are admissible if and only if \(\rho_G(S(g))/\rho_E(g)\) \(\leq\kappa\), where \(S= T+ T'\), \[ T(g)(x)= \int^\infty_x u^{(\alpha/n)-1)} g(u)\,du,\;T'(g)(x)= x^{(\alpha/n)-1} \int^x_0 g(u)\,du. \]
    0 references
    0 references
    0 references
    0 references
    0 references
    Riesz potential operator
    0 references
    rearrangement invariant function spaces
    0 references
    real interpolation
    0 references
    non-increasing rearrangements
    0 references