Stabbing simplices by points and flats

Abstract

The following result was proved by Barany in 1982: For every d >= 1 there exists c_d > 0 such that for every n-point set S in R^d there is a point p in R^d contained in at least c_d n^{d+1} - O(n^d) of the simplices spanned by S. We investigate the largest possible value of c_d. It was known that c_d <= 1/(2^d(d+1)!) (this estimate actually holds for every point set S). We construct sets showing that c_d <= (d+1)^{-(d+1)}, and we conjecture this estimate to be tight. The best known lower bound, due to Wagner, is c_d >= gamma_d := (d^2+1)/((d+1)!(d+1)^{d+1}); in his method, p can be chosen as any centerpoint of S. We construct n-point sets with a centerpoint that is contained in no more than gamma_d n^{d+1}+O(n^d) simplices spanned by S, thus showing that the approach using an arbitrary centerpoint cannot be further improved. We also prove that for every n-point set S in R^d there exists a (d-2)-flat that stabs at least c_{d,d-2} n^3 - O(n^2) of the triangles spanned by S, with c_{d,d-2}>=(1/24)(1- 1/(2d-1)^2). To this end, we establish an equipartition result of independent interest (generalizing planar results of Buck and Buck and of Ceder): Every mass distribution in R^d can be divided into 4d-2 equal parts by 2d-1 hyperplanes intersecting in a common (d-2)-flat.