Uniform estimates of nonlinear spectral gaps (Q497347): Difference between revisions

Nonlinear spectral gaps play an important role both in geometric group theory and metric geometry. Throughout the paper graphs and metric spaces are always assumed to contain at least two distinct points. The authors' aim of this paper is to establish a more accurate lower estimate of the nonlinear spectral gap of a given graph, which is independent of target metric spaces, by generalizing the path-method developed by \textit{M. Jerrum} and \textit{A. Sinclair} [SIAM J. Comput. 18, No. 6, 1149--1178 (1989; Zbl 0723.05107)], \textit{P. Diaconis} and \textit{D. Stroock} [Ann. Appl. Probab. 1, No. 1, 36--61 (1991; Zbl 0731.60061)], \textit{J. Quastel} [Commun. Pure Appl. Math. 45, No. 6, 623--679 (1992; Zbl 0769.60097)], \textit{P. Diaconis} and \textit{L. Saloff-Coste} [Ann. Appl. Probab. 3, No. 3, 696--730 (1993; Zbl 0799.60058)] and others. Although the original path method is well-known in the context of random walks as one giving only rough estimates of the linear spectral gap, they observe that their generalization gives asymptotically sharp estimates of the nonlinear spectral gaps in some examples. The authors, generalizing the path method, show that nonlinear spectral gaps of a finite connected graph are uniformly bounded from below by a positive constant which is independent of the target metric space. They also apply their result to an \(r\)-ball \(T_{d,r}\) in the \(d\)-regular tree, and observe that the asymptotic behavior of nonlinear spectral gaps of \(T_{d,r}\) as \(r\rightarrow\infty\) does not depend on the target metric space, which is in contrast to the case of a sequence of expanderas. Finally, the authors consider the \(n\)-dimensional Hamming cube \(H_{n}\) and obtain an estimate of its nonlinear spectral gap with respect to an arbitrary metric space, which is asymptotically sharp as \(n\rightarrow\infty\) This paper is a good article for researchers working on nonlinear spectral gaps.