A Bernstein theorem for affine maximal-type hypersurfaces (Q1713492): Difference between revisions

Consider the graph \(\mathcal{M}_{\Omega} = \{(x,z)\in\mathbb{R}^{N+1}:z=u(x), \ x\in\Omega\}\), where \(u\) is a convex function over \(\Omega\subseteq\mathbb{R}^N\). The affine maximal-type equation \[ D_{ij}(U^{ij}w) = 0,\] where \[w = [\mathrm{det}(D^2u)]^{-\theta}\] is equivalent to \[ u^{ij}D_{ij}w = 0,\] where \([u^{ij}]\) is the inverse matrix of the metric \([u_{ij}]\) of the graph \(M_{\Omega}\). \newline The authors prove that for \(N=2\) and \(\theta \neq 0,1/2\), if the graph of \(u\) is complete with respect to a \(\theta'\)-affine metric for some \(\theta'\) satisfying \((2\theta-1)\theta'\geq 1/8\), then it must be a paraboloid. In the case \(N\geq 2\), they prove that if \[ 0 \neq \theta \in \left(-\infty,\frac{1}{2}-\frac{N-1}{4N}\sqrt{N}\right)\bigcup\left(\frac{1}{2}+\frac{N-1}{4N}\sqrt{N},+\infty\right),\] and the Calabi metric is complete, then the convex entire solution of the equation above must be a quadratic function. As a corollary, they obtain that if the Calabi metric is complete and \(\theta = \frac{N+1}{N+2}\) for \(2\leq N \leq 3\), or \(\theta = 1\) for \(2 \leq N \leq 5\), then the entire graph of convex solutions to the equation above must be a paraboloid.