Max - Planck - Institut für Mathematik in den Naturwissenschaften Leipzig Ollivier ’ s Ricci curvature , local clustering and curvature

In Riemannian geometry, Ricci curvature controls how fast geodesics emanating from a common source are diverging on average, or equivalently, how fast the volume of distance balls grows as a function of the radius. Recently, such ideas have been extended to Markov processes and metric spaces. Employing a definition of generalized Ricci curvature proposed by Ollivier and applied in graph theory by Lin-Yau, we derive lower Ricci curvature bounds on graphs in terms of local clustering coefficients, that is, the relative proportion of connected neighbors among all the neighbors of a vertex. This translates the above Riemannian ideas into a combinatorial setting. We also study curvature dimension inequalities on graphs, building upon previous work of several authors.