A sufficient condition for a bounded set of positive Lebesgue measure in \(\mathbb{R}^2\) or \(\mathbb{R}^3\) to contain its centroid (Q1704428)

The authors provide sharp sufficient conditions for a subset of \({\mathbb R}^2\) and \({\mathbb R}^3\) to contain its centroid. They are: If \(U\) is a bounded, Lebesgue-measurable subset of \({\mathbb R}^2\) (respectively, of \({\mathbb R}^3\)) with positive measure, and if the largest open ball contained in \(U\) has radius \(r_1\) and the smallest closed ball containing \(U\) has radius \(r_2\), then \(U\) contains its centroid if \(r_2/r_1 < a_2\), where \(a_2\approx 1.82001\) is the unique positive zero of a certain rather complicated function of one variable (respectively, if \(r_2/r_1 < a_3\), where \(a_3\approx 1.71667\) is the unique positive real root of the polynomial \(f_3(x) = x^4 - 2x^3 + 2x - 2\)). The result is ``sharp'' in the sense that there exist bounded Lebesgue-measurable \(U_2\subset {\mathbb R}^2\) and \(U_3\subset {\mathbb R}^3\), which do not contain their centroids, although the ratio of the radii of the smallest closed ball containing \(U_2\) (respectively \(U_3\)) and the largest open ball contained in \(U_2\) (respectively \(U_3\)) is equal to \(a_2\) (respectively \(a_3\)).