Skip to content
Preprint

Power Concavity and Log-Concavity for Large-Parameter Robin 2-Hessian Problems

Oct 2026 · 0 citations · 21 references
Mathematics

Abstract

Let $\Omega\subset\mathbb R^n$, $n\ge3$, be a bounded uniformly convex $C^{3,\gamma}$ domain, where $0<\gamma<1$. We study the limit as the positive Robin parameter $\alpha\to\infty$. For the $2$-Hessian torsion problem, let $U$ be the admissible Dirichlet solution. For all sufficiently large $\alpha$, we obtain a unique admissible Robin solution $u_\alpha$ and establish the optimal-order estimate \[ \|u_\alpha-U\|_{C^{2,\gamma}(\overline\Omega)} \le C\alpha^{-1}. \] Moreover, $-\sqrt{-u_\alpha}$ has a uniform positive lower bound on its Hessian, and hence $\sqrt{-u_\alpha}$ is strictly concave. For the normalized eigenvalue problem, let $(\lambda_D,\varphi_D)$ be the normalized negative admissible Dirichlet eigenpair, and assume $\varphi_D\in C^{2,\gamma}(\overline\Omega)$. For all sufficiently large $\alpha$, there is a unique normalized negative admissible Robin eigenpair $(\lambda_\alpha,\varphi_\alpha)$, which satisfies \[ \|\varphi_\alpha-\varphi_D\|_{C^{2,\gamma}(\overline\Omega)} +|\lambda_\alpha-\lambda_D|\le C\alpha^{-1}. \] The Hessian of $-\log(-\varphi_\alpha)$ also has a uniform positive lower bound. The main analytic tool is a uniform H\"older estimate for linear Robin problems, with an $\alpha^{-1}$ gain for $C^{2,\gamma}$ boundary data. A normalized augmented inverse removes the scaling kernel in the eigenvalue problem. A Schur-complement argument combines the $C^2$ convergence with boundary geometry to prove convexity up to the boundary and anisotropic Hessian lower bounds at the scale ${\rm dist}(x,\partial\Omega)+\alpha^{-1}$. The convergence rate $O(\alpha^{-1})$ for the Robin solution functions is optimal even in $C^0$, as follows from their boundary conditions.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.