New quermassintegral and Poincar\'{e} type inequalities for non-convex domains

TL;DR

This paper introduces new inequalities for non-convex domains by studying a specific inverse curvature flow in Euclidean space, demonstrating convergence to spheres, and deriving Poincaré-type inequalities for k-convex hypersurfaces.

Contribution

It presents novel inverse curvature flow results for non-convex domains and establishes new geometric inequalities, extending classical results to broader classes of domains.

Findings

Flow exists for all time and converges to a sphere

New Alexandrov-Fenchel-type inequalities for non-convex domains

Poincaré-type inequality for k-convex hypersurfaces

Abstract

In the first part of this paper, we study the following non-homogeneous, locally constrained inverse curvature flow in Euclidean space $\mathbb{R}^{n+1}$, \begin{align*} \dot{x}=\left(\frac{1}{\frac{E_k(\hat{\kappa})}{E_{k-1}(\hat{\kappa})}-\alpha }-\langle x,\nu\rangle\right)\nu, \quad k=2,3,\ldots,n-1. \end{align*} Assuming that the initial hypersurface $\mathcal{M}_0 \subset \mathbb{R}^{n+1}$ is star-shaped and its shifted principal curvatures $\hat{\kappa}=\kappa+\alpha(1,\ldots,1)$ lie in the convex set \begin{align*} \Gamma_{\alpha,k}:=\Gamma_{k-1}\cap \{\lambda\in \mathbb{R}^n:\, E_k(\lambda)-\alpha E_{k-1}(\lambda)>0\}, \end{align*} we show that the flow admits a smooth solution that exists for all positive times, and it converges smoothly to a round sphere. As a corollary, we obtain a new set of Alexandrov-Fenchel-type inequalities for non-convex domains. In the second part, we derive a Poincar\'{e} type inequality for $k$-convex hypersurfaces which complements a more general version of the well-known Heintze-Karcher inequality.