Gradient estimates, Bakry-Emery Ricci curvature and ellipticity for unbounded graph Laplacians (Q6651625)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Gradient estimates, Bakry-Emery Ricci curvature and ellipticity for unbounded graph Laplacians |
scientific article; zbMATH DE number 7956687
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Gradient estimates, Bakry-Emery Ricci curvature and ellipticity for unbounded graph Laplacians |
scientific article; zbMATH DE number 7956687 |
Statements
Gradient estimates, Bakry-Emery Ricci curvature and ellipticity for unbounded graph Laplacians (English)
0 references
10 December 2024
0 references
It has been known for many decades that there is a deep relationship between Ricci curvature and the heat equation. In particular, lower Ricci curvature bounds can be characterized via gradient estimates for the heat semigroup. Recently, there has been remarkable interest in discrete versions of Ricci curvature.\N\NThe main result of the paper under review is a gradient estimate under Bakry-Émery curvature bounds for unbounded graph Laplacians which satisfy an ellipticity assumption. As applications, the authors investigate completeness and finiteness of stochastically complete graphs under Bakry-Émery curvature bounds.\N\NThe paper is structured as follows. In Section 2, the authors introduce the basic notions. In Section 3 they prove a Green's formula which is a consequence of the ellipticity condition (EC). In Section 4 they compute the derivative of the function \(s \mapsto \langle \Gamma_k(P_{t-s}f),P_s1_x\rangle\) with respect to \(s\) under very mild conditions. Here, \(P_t=e^{-tL}\) for the graph Laplacian \(L\) and \(\Gamma_k\) is defined (in Section 2) iteratively via the Gamma calculus. They put these two pieces together in Section 5 to prove the main result. They also investigate the applications of the main theorem for stochastically complete graphs. Finally, they provide examples of graphs for which the new method gives gradient estimates, but all previous approaches fail to do so.\N\NThis is an paper of interdisciplinary nature, useful for current developments.
0 references
discrete Ricci curvature
0 references
graph Laplacians
0 references