Stability results for the volume of random simplices (Q2879436)

Let \(K\) be a convex body in \({\mathbb R}^d\). \(V_d\) stands for the \(d\)-dimensional volume, the convex hull of the points \(x_1, x_2,\dots x_n\) is denoted by \([x_1, x_2,\dots,x_n]\), and \(\gamma(K)\) is the centroid of \(K\). For any \(n\geq d+1\) and \(p>0\), let \(E_n^p(K)= [V_d(K)]^{(-n-p)} \int_K\dots \int_K V_d([x_1,\dots, x_n])^p\, dx_1\dots dx_n\), and for a fixed \(x\in {\mathbb R}^d\), let \(E_x^p(K)= [V_d(K)]^{(-n-p)} \int_K\dots \int_K V_d([x, x_1,\dots, x_n])^p\, dx_1\dots dx_d\). Specifically, when \(x=\gamma(K)\) denote \(E_x^p(K)\) by \(E_*^p(K)\). In particular, for integer \(p\), \(E_{d+1}^p(K)\) is the expectation of the \(p\)th moment of the relative volume of simplices in \(K\). Clearly, \(E_n^p(K)\) and \(E_*^p(K)\) are invariant under non-singular affine transformations, and \(E_o^p(K)\) is invariant under non-singular linear transformations, where \(o\) stands for the origin. Note that for fixed \(K\) and \(p\geq 1\), \(E_x^p(K)\) is a strictly convex function of \(x\), therefore it attains its minimum at a unique point. If \(K\) is \(o\)-symmetric, then the minimum is attained at \(o\), and \(E_o^p(K) = E_*^p(K)\). In the rest of Section 1 the authors give an overview of the history of the quantities \(E_n^p(K)\) and \(E_x^p(K)\) and their various connections. In particular, for a convex body \(K\) of volume 1, the expected volume of random simplices in \(K\) is minimized if \(K\) is an ellipsoid, and for \(d=2\), maximized, if \(K\) is a triangle. The main results are presented in Section 2, whose proofs are found in the subsequent parts. The main aim is to provide stability versions of classical theorems of Stochastic Geometry. Section 6 contains further corollaries.