Gradient estimate for a locally constrained flow in $\mathbb H^{n+1}$

Problem's Description

The following hypersurface flow in $\mathbb H^{n+1}$ was introduced by Brendle-Guan-Li: \begin{align} X_t=\Big(\cosh(\rho) \frac{\sigma_{k}}{\sigma_{k+1}}(\kappa)-\frac{u}{c_{n,k}}\Big)\nu, \label{flow} \end{align} where $\nu$ the outer normal, $\kappa$ the principal curvature vector. The flow preserves $k$-th quermassintegral $\mathcal{A}_{k}$ and decreases $k+1)$-th quermassintegral $\mathcal{A}_{k+1}$. All regularity estimates and convergence would follow if one can establish the gradient estimate (or preservation of starshapedness) for solution of the flow. In turn, Alexandrov-Fenchel inequality for quermassintegrals in hyperbolic space can be established for general starshaped $k+1$-domains.

When $k=0$, Alexandrov-Fenchel inequality is the Minkowski inequality. This inequality is proved by Brendle-Guan-Li for general mean convex starshaped domains in $\mathbb H^{n+1}$. Alexandrov-Fenchel inequality is known for horo-convex domains (Wang-Xia).

\medskip

\noindent {\bf Open problem}: prove the gradient estimate for flow (\ref{flow}) $\forall k< n-1$.

\medskip

The case $k=n-1$ is known, as the flow preserved convexity.

1. Article Isoperimetric type problem and hypersurface flows

   • Pengfei Guan and Junfang Li

   Journal of Mathematical Study 56 (1), 56-80, 2021