Marcinkiewicz integrals with non-smooth kernels on spaces of homogeneous type (Q343252)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Marcinkiewicz integrals with non-smooth kernels on spaces of homogeneous type
scientific article

    Statements

    Marcinkiewicz integrals with non-smooth kernels on spaces of homogeneous type (English)
    0 references
    0 references
    25 November 2016
    0 references
    Let \((X,d,\mu)\) be a space of homogeneous type, endowed with a quasi-metric \(d\) on \(X\) satisfying \(d(x,z)\leq \kappa(d(x,y)+d(y,z))\) for some \(\kappa>0\) and for all \(x,y,z\in X\), and a regular Borel measure \(\mu\) on \(X\) with the doubling property \(\mu(B(x;2r))\leq C \mu(B(x;r))<\infty\) for some fixed constant \(C\geq1\), for all \(x\in X\) and for all \(r>0\), where \(B(x,r)=\{y\in X; d(x,y)<r\}\). From the definition, it follows that there exist \(C_0\geq 1\) and \(n>0\) such that \(\mu(B(x;\lambda r))\leq C_0\lambda^n \mu(B(x;r))\). There also exist \(C_1\geq1\) and \(0\leq N\leq n\) such that \(\mu(B(y;r))\leq C_1 (1+d(x,y)/r)^N\mu(B(x;r))\) for \(x,y\in X\) and \(r>0\). The author considers a class of integral operators \(\{A_t\}_{t>0}\), which are a sort of approximations of the identity (introduced by Duong and McIntosh): \(A_tf(x)=\int_{X}a_t(x,y)f(y)d\mu(y)\), and \(a_t(x,y)\) satisfy the condition \((*)\;| a_t(x,y)| \leq h_t(x,y)=\bigl(\mu(B(x;t^{1/m})\bigr)^{-1} s\bigl((d(x,y))^mt^{-1} \bigr)\), for some positive constant \(m\) and a positive, bounded and decreasing function \(s\) satisfying \(\lim_{r\to\infty}r^{(n+\sigma)}s(r^m)=0\) for some \(\sigma>N\). Now, Marcinkiewicz interal on \(X\) is defined as follows: For a suitable function \(f\) define the linear operator \(F(f)\) by \(F(f)(x,r)=\int_{d(x,y)<r}K(x,y)f(y)d\mu(y)\), where \(K(x,y)\) is a measurable function on \(X\times X\setminus \{(x,x); x\in X\}\). Define the Marcinkiewicz integral \(\nu(f)\) by \(\nu(f)(x)=\bigl(\int_0^\infty| F(f)(x,r)| ^2dr/r^3\bigr)^{1/2}\). The author shows that \(\nu\) is \(L^p\) bounded \((1<p<\infty)\) and weak \((1,1)\) bounded if the following three conditions (a), (b) and (c). (a) \(\nu\) is \(L^r\) bounded for some \(1<r<\infty\); (b) There exist \(c_2, C_2>0\) such that \(\int_{d(x,y)\geq c_2t^{1/m}}| K(x,y)-K_t(x,y)| d(x,y)^{-1}d\mu(y)\leq C_2\) for all \(y\in X\); (c) There exist \(c_3, C_3, \beta>0\) such that \(| K(x,y)-K^t(x,y)| d(x,y)^{-1} \leq C_3\mu(B(x;d(x,y)))^{-1}(t^{1/m}/d(x,y))^\beta\) whenever \(d(x,y)\geq c_3t^{1/m}\). Here, \(K^t(x,y)\) represents the kernel of the operator \(A_tF\), and \(K_t(x,y)\) represents the kernel of the operator \(FB_t\) for another class of operators similar to \(A_t\). Under similar appropriate conditions, the same \(L^p\) \((1<p<\infty)\) and weak \((1,1)\) boundedness for singular integrals were given by \textit{T.D. Xuan} and \textit{A. MacIntosh} [Rev. Mat. Iberoam. 15, No. 2, 233--265 (1999; Zbl 0980.42007)]. The author gives another sufficient conditions giving \(L^p\) \((1<p<\infty)\) and weak \((1,1)\) boundedness for \(\nu\). He also discusses \(L^p\) \((1<p<\infty)\) boundedness of commutators of BMO functions and \(\nu\).
    0 references
    0 references
    0 references
    Marcinkiewicz integral
    0 references
    non-smooth kernel
    0 references
    \(L^p\) estimate
    0 references
    weak \((1,1)\) estimate
    0 references
    spaces of homogeneous type
    0 references
    commutators
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references