Stochastic reverse isoperimetric inequalities in the plane (Q2665241)

Many questions in geometry study measurements of convex body, e.g. its (Euklidean) volume. One of them is the volume of Mahler: For a symmetric convex body \(\mathcal{K}\) the inner product of the embedding Euclidean space defines the polar body \(\mathcal{K}^\circ\). Then the volume of Mahler is the product of the volumes of \(\mathcal{K}\) and \(\mathcal{K}^\circ\), i.e. \( P(K) := |\mathcal{K}| \cdot |\mathcal{K}^\circ|. \) For the two-dimensional Euclidean space the upper bound of this volume is known to be the Euclidean ball (Inequality of Blaschke-Santalo). On the other side, lower bounds are only known for some particular cases (see Mahler's conjecture). In recent years, it has been shown that some classical inequalities follow from a local stochastic dominance for naturally associated random polytopes. A typical example of such random sets is the convex hull of the columns of a random matrix; the expectation of the volume of this polytope is maximized by \(N\) independent random vectors uniformly distributed in the Euclidean ball of volume one. In this note the author improves some isoperimetric inequalities by attaching a stochastic model to some classical inequalities, such as Mahler's Theorem, and a reverse Lutwak-Zhang inequality concerning the polar for \(L_p\) centroid bodies. Furthermore he obtains also a dual counterpart to a result of Bisztriczky-Böröczky concerning centroid bodies. All these inequalities concern central symmetric bodies together with their polars and suiting symmetrizations. The stated theorems give upper bounds of the expectation of randomly choosen convex bodies -- and where the random bodies should be taken from. The proofs use the shadow system of a bounded point set. This system is a family of convex sets given by a point set and a given direction. It defines a convex function that can be used to solve isoperimetric type problems.