On automorphisms of complements of analytic subsets in \({\mathbb{C}}^ n\) (Q914005)

The paper gives several interesting results on the automorphism group \(Aut_{{\mathcal O}}(X)\) of the complementary X in \({\mathbb{C}}^ n\) of an analytic set A. He calls X \(Aut_{{\mathcal O}}\)-homogeneous (resp. G- homogeneous) when X is homogeneous under \(Aut_{{\mathcal O}}(X)\) (resp. under a Lie group (real or complex) in \(Aut_{{\mathcal O}}(X))\). W. Kaup has given an example with a discute such that X is \(Aut_{{\mathcal O}}\)- homogeneous and not G-homogeneous. W. Rudin and J. L. Rosay have given another example with A discrete such that \(Aut_{{\mathcal O}}(X)=\{id\}\). Among others, the author proves the following: 1) A algebraic and \(co\dim_{{\mathbb{C}}}(A)\geq 2\). Then X is \(Aut_{{\mathcal O}}\) homogeneous and if the number of two dimensional components is at least two, then X is not G-homogeneous. 2) Let be \(A=\{(z,w)\in {\mathbb{C}}^ 2|\) \(z.w=1| \}\). Then \({\mathbb{C}}^ 2-A\) is \(Aut_{{\mathcal O}}\)-homogeneous and not G-homogeneous. 3) X admits a non-constant holomorphic map to a hyperbolic manifold. Then X is not \(Aut_{{\mathcal O}}\)-homogeneous. 4) A is an algebraic hypersurface containing \(n+1\) hyperplanes in general position, then \(Aut_{{\mathcal O}}(X)\) is finite. 5) \(X={\mathbb{C}}^*\times {\mathbb{C}}^*-\{(1,1),(2,1),(1,3)\}\). Then the identity is the only algebraic automorphism on this quasi affine variety, but X is \(Aut_{{\mathcal O}}\)-homogeneous. Let us give some indications on the proofs. 1) Let be \(\Pi_ i\) the i-canonical projection of \({\mathbb{C}}^ n\) on \({\mathbb{C}}^{n-1}\). By considering automorphisms of the form \(z\to (z_ 1,...,z_{i-1},f(\Pi_ i(z)),z_{i+1},...,z_ n)\) and using Chevalley's theorem on constructible sets, one has that \(U={\mathbb{C}}^ n\setminus (UF_ i)\) is a non empty set contained in an open H-orbit. (H: group of algebraic transformation leaving A pointwise fixed.) Then using the canonical projection \({\mathbb{C}}^ n-\{x\}\to {\mathbb{P}}_{n-1}\) and Chevalley's theorem once again, he proves the beginning of the proposition by taking adapted coordinates. For X not G-homogeneous, he uses a general theorem of Borel on the ends of homogeneous manifolds under a Lie group. 2) He gives explicitly automorphisms and uses also the theorem of Borel. 3) Uses results on Kobayashi metrics. An example is the following: Let be \(f\in {\mathcal O}({\mathbb{C}}^ n)\) a non-constant holomorphic function. Then if \(X=\{z\in {\mathbb{C}}^ n|\) f(z)\(\neq 0\) and f(z)\(\neq 1\}\), X is not \(Aut_{{\mathcal O}}\)-homogeneous (\({\mathbb{C}}-\{0,1\}\) is hyperbolic\(\}\). 4) Uses the theory of the logarithmic Kodaira-dimension for non complete varieties. 5) He proves explicitly that X is \(Aut_{{\mathcal O}}\)-homogeneous and uses that an algebraic morphism is of the form \((z,w)\to (\lambda z^ kw^ n\), \(\mu z^ pw^ q)\).