On the Lichnerowicz estimate for Laplacians on graphs under spectral curvature assumptions (Q2090549)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On the Lichnerowicz estimate for Laplacians on graphs under spectral curvature assumptions
scientific article

    Statements

    On the Lichnerowicz estimate for Laplacians on graphs under spectral curvature assumptions (English)
    0 references
    0 references
    25 October 2022
    0 references
    A triple \(G=(V,w,m)\) that consists of a finite vertex set \(V\), a symmetric function \(w: V\times V\rightarrow [0,\infty)\) which is zero on diagonal, and a vertex measure \(m: V\rightarrow (0,\infty)\) is called a finite weighted and measured graph. Let \(C(V)\) denote the set of real valued functions defined on \(V\). On the graph \(G\), the discrete Laplacian \(\Delta: C(V)\rightarrow C(V)\) is given in the following way \[ \Delta f(x):=\sum_{y\in V} \frac{w(x,y)}{m(x)} (f(y)-f(x)),\quad x\in V. \] Let \(K>0\), \(n\in(0,\infty]\) and \(\rho\in C(V)\). It is said that the \((K,n,\rho)\)-spectrally positive Ricci curvature dimension condition on \(G\) is satisfied if and only if the following two conditions are satisfied: \par i) the variable curvature dimension condition \(CD(\rho,n)\) holds \[ \Gamma_2(f)(x)\geq \rho(x)\Gamma(f)(x)+\frac{1}{n}(\Delta f)^2(x),\quad f\in C(V), \quad x\in V. \] ii) as well as \[ -\frac{1}{2}\langle -\Delta f,f\rangle+\langle \rho f,f\rangle\geq K\langle f,f\rangle,\quad f\in C(V). \] The following eigenvalue estimate is established and proved. Let \(K>0\), \(n\in(1,\infty]\) and \(\rho\in C(V)\). Let \(G=(V,w,m)\) be a finite weighted and measured graph on which the \((K,n,\rho)\)-spectrally positive Ricci curvature dimension condition is satisfied. Then, \[ \lambda\geq \frac{n}{n-1}K, \] where \(\lambda=\lambda(G)\) is the first positive eigenvalue of \(\Delta\) in \(\ell^2(V,m)\).
    0 references
    finite weighted and measured graph
    0 references
    Ricci curvature dimension
    0 references
    eigenvalue estimate
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references