Total curvature and the isoperimetric inequality in Cartan-Hadamard manifolds

DOI10.1007/S12220-021-00801-2MaRDI QIDQ2074504zbMATH OpenOpenAlexFDO

Publication date 10 February 2022

Published in The Journal of Geometric Analysis (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1908.09814

convex hull cut locus inf-convolution isoperimetric profile medial axis positive reach quermassintegrals semiconcave functions signed distance function

Mathematics Subject Classification ID

Geometric measure and integration theory, integral and normal currents in optimization (49Q15) Length, area, volume and convex sets (aspects of convex geometry) (52A38) Global Riemannian geometry, including pinching (53C20) Elliptic equations on manifolds, general theory (58J05)

Abstract: We obtain an explicit formula for comparing total curvature of level sets of functions on Riemannian manifolds, and develop some applications of this result to the isoperimetric problem in spaces of nonpositive curvature.

Review text

A Cartan-Hadamard manifold \(M\) is a complete simply connected Riemannian space with nonpositive sectional curvature. In the present paper the authors are concerned with generalizing some fundamental properties of Euclidan space \(\mathbb{R}^{n}\) to Cartan-Hadamard manifolds. A closed convex hypersurface \(\Gamma \subset M\) is the boundary of a convex set with interior points. If \(\Gamma \) is of regularity class \(\mathcal{C}^{1,1}\), then its Gauss-Kronecket curvature \(GK\), i.e., the determinant of second fundamental form, is well defined almost everywhere. So the total curvature of \(\Gamma \) is defined as \(\mathcal{G}(\Gamma )=\int_{\Gamma }|GK|d\sigma \), where \(d\sigma \) is the volume form of \(\Gamma \). In the case that \(M=R^{n}\), it is easy to see that \[ \mathcal{G}(\Gamma )\geq\mathrm{vol}(\mathbb{S}^{n-1}). \tag{1} \] So the following interesting problem is posed: Does the total curvature inequality (1) hold for convex \(\mathcal{C}^{1,1}\) hypersurfaces in Cartan-Hadamard manifolds \(M^{n}\)? A prime motivation for studying the previous problem is its connection to the classical isoperimetric inequality which states that for any bounded set \(\Omega \subset \mathbb{R}^{n}\), \[ \mathrm{per}(\Omega )^{n}\geq \frac{\mathrm{per}(\mathbb{B}^{n})^{n}}{\mathrm{vol}(\mathbb{B}^{n})^{n-1} }\mathrm{vol}(\Omega )^{n-1}, \tag{2} \] where \(\mathrm{per}\) stands for perimeter, and \(\mathbb{B}^{n}\) is the unit ball in \(\mathbb{R}^{n}\), so \(\mathrm{per}(\mathbb{B}^{n})=\mathrm{vol}(\mathbb{S}^{n-1})\). Furthermore, equality holds if \(\Omega \) is a ball in \(\mathbb{R}^{n}\). So, another interesting problem is posed: Does the isoperimetric inequality (2) hold for bounded sets in Cartan-Hadamard manifolds \(M^{n}\)? The main theorem of the paper states that the total curvature inequality (1) implies the isoperimetric inequality (2) in all dimensions. Notice that this results is motivated by a theorem of \textit{B. Kleiner} [Invent. Math. 108, No. 1, 37--47 (1992; Zbl 0770.53031)] who showed the same theorem for \(n=3\).

Recommendations

Cites work

Cited in

(17)

This page was built for publication: Total curvature and the isoperimetric inequality in Cartan-Hadamard manifolds

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2074504)