A characterization of convex cones (Q1002587)

The author considers domains of an affine space in the projective category via the well-known equivariant embedding from \((\mathbb{R}^n,\text{Aff}(n,\mathbb{R}))\) into \((\mathbb{R}{\mathbf P}^n,\text{PGL}(n+1,\mathbb{R}))\). A domain \(\Omega\) of \(\mathbb{R}{\mathbf P}^n\), is called convex if there exists an affine space \(H\subset\mathbb{R}{\mathbf P}\), such that \(\Omega\) is a convex subset of \(H\). In the paper the following theorem is proved. Let \(\Omega\) be a convex domain with a face \(F\) of codimension 1. Then \(\Omega\) is a cone over \(F\) if the automorphism group of \(\Omega\) has an orbit accumulating at a point of \(F\). As a corollary the author proves: Let \(\Omega\) be a convex domain in \(\mathbb{R}^n\) and \(F\) an \((n-1)\)-dimensional face of \(\Omega\). Suppose that there is a sequence \(\{g_i\}\) of affine transformation which preserve \(\Omega\) and a point \(x\) in the interior of \(\Omega\) such that the sequence \(\{g_i(x)\}\) accumulates at a point of \(F\). Then \(\Omega= \mathbb{R}^+\times F\).