Nonlinear oblique boundary value problems for Hessian equations in two dimensions (Q1905028)

This long paper is devoted to the study of the following nonlinear oblique boundary value problem on a bounded domain \(\Omega\subset \mathbb{R}^2\): \[ F(D^2 u)= g(x, u, Du)\quad\text{in}\quad \Omega,\quad b(x, u, Du)= 0\quad\text{on} \quad \partial\Omega. \] The function \(F\) is given by \(F(D^2 u)= f(\lambda_1, \lambda_2)\), where \(\lambda_1\), \(\lambda_2\) denote the eigenvalues of the Hessian matrix \(D^2 u\) and \(f\) is a suitable symmetric function defined on an open, convex, symmetric (under interchange of \(\lambda_1\) and \(\lambda_2\)) region \(\Sigma\subset \mathbb{R}^2\). The solution \(u\in C^2(\Omega)\) is called admissible if the corresponding \((\lambda_1, \lambda_2)\in \Sigma\). The function \(f\) and the domain \(\Sigma\) are subject to various restrictions. The function \(g(x, z, p)\) is assumed to be smooth and nondecreasing with respect to \(z\). If \(\nu\) denotes the inner unit normal to \(\partial\Omega\), the case \(b(x, u, Du)= D_\nu u+ \phi(x, u)\) for suitable \(\phi\) is investigated as well as the case \(b(x, u, Du)= D_\beta u+ \phi(x, u)\), where \(\beta\) is a unit vector field on \(\partial\Omega\) with \(\beta\cdot \nu> 0\) and satisfying a special further condition on the boundary. A typical result is the following: Under suitable hypotheses on \(\Sigma\), \(f\), \(\Omega\), \(g\), \(\phi\) and \(\beta\), the boundary value problem \(F(D^2 u)= g(x, u)\) in \(\Omega\), \(D_\beta u+ \phi(x, u)= 0\) on \(\partial\Omega\) has a unique admissible solution \(u\in C^{2, \alpha}(\Omega)\) for some \(\alpha\in (0, 1)\). When \(g\) depends also on \(Du\), the above result is proved provided \(f\) and \(g\) satisfy suitable additional restrictions and assuming the existence of a subsolution. Furthermore, the fully nolinear boundary condition \(D_\nu u+ \phi(x, u, \delta u)= 0\) on \(\partial \Omega\) is invstigated. Here \(\delta\) denotes the tangential gradient operator. To prove existence, the method of continuity is used, and sufficient conditions for the existence of subsolutions are given. To use the continuity method, the author proves a priori bounds for \(u\), \(Du\) and \(D^2 u\). Of course, this is the most significant part of the paper. Extensions to the degenerate situation are also discussed and interesting examples to justify some hypotheses are exhibited.