Stability properties of neighbourly random polytopes (Q1017918): Difference between revisions

The authors study some geometric properties of convex symmetric polytopes related to the structure of their faces. Let \(\nu(K)\) be the set of vertices of a convex symmetric polytope \(K\subset\mathbb{R}^k\), and, for \(1\leq m<k\), denote by \(\nu_m(K)\) the family of all nonempty subsets of \(\nu(K)\) with cardinality less than or equal to \(m\) and not containing a pair of antipodal vertices. For \(2\leq m<k\), the convex symmetric polytope \(K\) is said to be \(m\)-neighbourly if and only if every \(V\in\nu_m(K)\) is a vertex set for a face of \(K\), which means that \(\text{conv}V\cap\text{conv}\{\nu(K)\setminus V\}=\emptyset\) for every \(V\in\nu_m(K)\). By a standard compactness argument this implies that there exists \(\alpha>1\) such that \(\text{conv}V\cap\alpha\text{conv}\{\nu(K)\setminus V\}=\emptyset\) for every \(V\in\nu_m(K)\). Motivated by the above, the authors introduce the notion of \(m\)-neighbourliness of \(K\) with parameter \(\alpha\) (where \(m=1\) is also allowed), shortly called \((m,\alpha)\)-neighbourliness. Further, \(\text{nei}_m(K)\) denotes the set \(\sup\{\alpha\mid K\text{ is }(m,\alpha)\)-neighbourly