Riesz means on graphs and discrete groups (Q1935431)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Riesz means on graphs and discrete groups
scientific article

    Statements

    Riesz means on graphs and discrete groups (English)
    0 references
    15 February 2013
    0 references
    For \(\alpha>0\) and \(R>0\), the Riesz mean of order \(\alpha\) is the operator defined by \[ m_{\alpha,R}(\Delta) = \int_0^2 m_{\alpha,R}(\lambda) \mathrm{d}E_\lambda \] where \(\Delta\) is the discrete Laplacian, \(\mathrm{d}E_\lambda\) is its spectral measure (so that \(\Delta = \int_0^2 \lambda \mathrm{d}E_\lambda\)) and \[ m_{\alpha,R}(\lambda) = \bigg( 1- \frac{\lambda}{R} \bigg)^\alpha_+. \] The main result concerns graphs satisfying the doubling volume property, {i.e.} there is a \(c\) such that \(\forall x \in \Gamma\) and \(r>0\), \(|B(x,2r)| \leq c |B(x,r)|\), an exponent \(D\) (similar to the polynomial growth exponent). See also [\textit{E. Russ}, Math. Scand. 87, No. 1, 133--160 (2000; Zbl 1008.60085)]. The main result is Theorem 1: {\parindent=6mm \begin{itemize}\item[(i)] If \(\alpha > D/2\), then \(m_{\alpha,R} (\Delta)\) are bounded on \(L^p\) , for \(p \in [1, \infty]\). \item[(ii)] If \(0 < \alpha < D/2\), then \(m_{\alpha,R}(\Delta)\) are bounded on \(L^p\) , for \(p \in (1, \infty)\), provided that \(\alpha > D| \frac{1}{p} - \frac{1}{2} |\). \end{itemize}} (As the authors note, the critical power for the Laplace operator on \(R^D\) in the above result is \((D - 1)/2\) rather than \(D/2\).) This result is not covered by the Hörmander multiplier theorem in the cases \(p = 1\) and \(p = \infty\) when \(\alpha > D/2\), as well as the case \(\alpha \in ( D| \frac{1}{p} - \frac{1}{2} |, D/2)\). This theorem has been proved for Riemannian manifolds \(M\) with \(\mathrm{Ric}(M) \geq 0\) and Lie groups of polynomial volume growth by \textit{G. Alexopoulos} and \textit{N. Lohoué} [Bull. Soc. Math. Fr. 122, No. 2, 209--223 (1994; Zbl 0832.22014)]. The proof involves decomposing \(m_{\alpha,R} (\lambda)\) into compactly supported components and estimating the corresponding kernels.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Riesz means
    0 references
    graphs
    0 references
    discrete Laplacian
    0 references
    discrete groups
    0 references
    Markov kernels
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references