Characterization of linearly convex domains in \(\mathbb C^n\) by their noncompact automorphism groups (Q842295)

By definition a domain \(\Omega\subset\mathbb C^n\) is linearly convex near a boundary point \(p\in\partial\Omega\) if there is a neighborhood \(U\) of \(p\) such that \((z+T_z^{10}\partial\Omega)\cap(\Omega\cap U =\emptyset\) for every \(z\in U\cap\partial\Omega\). The author extends a result of \textit{H. Gaussier} [Mich. Math. J. 44, No. 2, 375--388 (1997; Zbl 0889.32032)] on convex domains in the following manner. Assume \(\partial\Omega\) is smooth linearly convex and of finite type \(2m\) near a point \(p_\infty\in\partial\Omega\) where \(p_\infty = \lim_{k\to\infty} F_k(p_0)\) for some inner point \(p_0\in\Omega\) and a suitable sequence \(F_1,F_2,\dots\) of biholomorphic automorphisms of \(\Omega\). Then necessarily \(\Omega\) is biholomorphically equivalent to a rigid polynomial domain of the form \(\{ z\in\mathbb C :\;\text{Re}\, z_1 + P(z_2,\dots,z_n) < 0\}\) with a real-valued non-degenerate plurisubharmonic polynomial \(P\) of degree \(\leq 2m\) on \(\mathbb C^{n-1}\). The proof is based on a scaling method with Carathéodory kernel convergence and borrows some ideas from M. Conrad's thesis associating polydiscs to linearly convex domains quoted in details in the paper.