The surface area deviation of the Euclidean ball and a polytope (Q1745266)

The authors study polytopal approximation of the Euclidean ball \(B^n_2\), with the special features that the approximating polytopes need not contain, nor be contained in, the ball and that the deviation is measured by the symmetric surface area difference. The latter is defined, for convex bodies \(K,L\) in \({\mathbb R}^n\), by \[ \Delta_s(K,L) = \text{vol}_{n-1}(\partial(K \cup L)) - \text{vol}_{n-1}(\partial(K \cap L)). \] The first theorem says that there is an absolute constant \(c>0\) such that for every integer \(n\geq 3\), there is an \(N_n\) such that for every \(N\geq N_n\) there is a polytope \(P_N\) in \({\mathbb R}^n\) with \(N\) vertices such that \[ \Delta_s(B_2^n,P_N) \leq c\frac{\text{vol}_{n-1}(\partial B_2^n)}{N^\frac{2}{n-1}}. \] The proof of this theorem is based on a probabilistic construction. The second theorem says that there are constants \(c>0\) and \(N_0 \in {\mathbb N}\) such that for all \(n \in {\mathbb N}\) with \(n \geq 2\), all \(N \in {\mathbb N}\) with \(N \geq N_0\) and all polytopes \(P_N\) in \({\mathbb R}^n\) with no more than \(N\) facets, the inequality above holds with the reversed inequality sign. As a corollary, this yields a lower bound for the approximation of the ball by simple polytopes with at most \(N\) vertices.