Davies-Gaffney-Grigor'yan Lemma on graphs (Q265097)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Davies-Gaffney-Grigor'yan Lemma on graphs
scientific article

    Statements

    Davies-Gaffney-Grigor'yan Lemma on graphs (English)
    0 references
    0 references
    0 references
    0 references
    1 April 2016
    0 references
    Let \((M,g)\) be a complete Riemannian manifold. Let \(p_t(x,y)\) be the minimal heat kernel on \(M\). Let \(\mu\) be the greatest lower bound on the \(L^2\) spectrum of the Laplacian on \(M\). The Davies-Gaffney-Grigoryan result can be stated as follows: Theorem DGG: If \(B_1\) and \(B_2\) are measurable subsets of \(M\) and if \(t>0\), then \[ \int_{B_1}\int_{B_2}p_t(x,y)\text{dvol}(x)\text{dvol}(y)\leq\{\text{vol}(B_1)\text{vol}(B_2)\}^{1/2}e^{-\mu t}e^{-\frac{d^2(B_1,B_2)}{4t}}. \] See [\textit{E. B. Davies}, J. Anal. Math. 58, 99--119 (1992; Zbl 0808.58041); \textit{M. P. Gaffney}, Commun. Pure Appl. Math. 12, 1--11 (1959; Zbl 0102.09202); \textit{A. A. Grigor'yan}, Mat. Sb. 182, No. 1, 55--87 (1991; Zbl 0743.58031); Proc. R. Soc. Edinb., Sect. A, Math. 124, No. 2, 353--362 (1994; Zbl 0812.58082)]. Although this result is known to fail in the context of graphs, the authors establish a version of this result for the continuous time heat kernel on graphs if the time \(t\) is large in comparison to the distance \(d(B_1,B_2)\) between \(B_1\) and \(B_2\). The authors use this estimate and the Li-Yau inequality on graphs [\textit{F. Bauer} et al., J. Differ. Geom. 99, No. 3, 359--405 (2015; Zbl 1323.35189)] to obtain strong heat kernel estimates for graphs which satisfy an exponential curvature dimension inequality. Section 1 contains an introduction to the problem, states the main results, and establishes notational conventions. In Section 2, the Li-Yau inequality on graphs is presented. The Yau Liouville Theorem, the Cheng Liouville Theorem, and the Cheng eigenvalue estimate on graphs are established. In Section 3, the extension of Theorem DGG to graphs is established. In Section 4, this extension is used to derive a new heat kernel estimate for finite graphs and to derive higher order eigenvalue estimates in terms of the distances between subsets of a finite graph.
    0 references
    Davies-Gaffney-Grigoryan Lemma
    0 references
    Li-Yau inequality on graphs
    0 references
    Harnack inequality
    0 references
    Yau Liouville Theorem
    0 references
    Cheng Liouville Theorem
    0 references
    Cheng eigenvalue estimate
    0 references
    finite graphs
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references