Applications of various inequalities to Minkowski geometry

DOI10.1007/BF01264921MaRDI QIDQ1801635zbMATH OpenOpenAlexFDO

Publication date 17 August 1993

Published in Geometriae Dedicata (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/bf01264921

Haar measure isoperimetrix Minkowski content Minkowski space self-surface-area

Minkowski geometries in nonlinear incidence geometry (51B20) Convex sets in (n) dimensions (including convex hypersurfaces) (52A20) Convexity and finite-dimensional Banach spaces (including special norms, zonoids, etc.) (aspects of convex geometry) (52A21) Mixed volumes and related topics in convex geometry (52A39) Inequalities and extremum problems involving convexity in convex geometry (52A40)

Review text

A finite-dimensional normed linear space \(X\) with a unit ball \(B\) is equipped with an auxiliary Euclidean structure which allows to identify \(X\) with some \(\mathbb{R}^ d\) and to normalize the Haar measure in each \(k\)- dimensional subspace of \(X\). In each subspace the Minkowski content \(\mu\) is a multiple of the appropriate Haar measure \(\lambda\). Here the author deals with the most interesting case \(k=d-1\) of hyperplanes of \(X\). Then \(\mu=\sigma(f)\lambda\), where \(f\) is the unit normal to \(H\), that is, \(f\) is a linear functional in the dual space \(X^*\) of \(X\) of Euclidean length 1 whose null space \(f^ \perp\) is \(H\). The author gives lower bounds for the self-surface-area of a ball in \(\mathbb{R}^ d\). More precisely, let \(\mu^{(1)}_ B\) and \(\mu^{(2)}_ B\) denote the area measures which arise from \(\sigma_ 1(f)= {\varepsilon_{d-1}\over \lambda(B\cap f^ \perp)}\) and \(\sigma_ 2(f)= {\lambda((B\cap f^ \perp)^ 0)\over \varepsilon_{d-1}}\) respectively, where \(\varepsilon_ d\) denotes the \(d\)-dimensional volume of a \(d\)- dimensional Euclidean unit ball and \(C^ 0\subseteq X^*\) is the dual of \(C\subseteq X\). Then \(\mu^{(1)}_ B(\partial B)\geq (d\varepsilon_ d)\Bigl({m_ d\over \varepsilon^ 2_ d}\Bigr)^{1/d}\) and \(\mu^{(2)}_ B(\partial B)\geq\Bigl({d\over\varepsilon_ d}\Bigr)\Bigl({4^ d\over d!}\Bigr)^{1/d}(m_ d)^{(d-1)/d}\), where \(m_ d\) is the minimum of \(\lambda(K)\lambda(K^ 0)\) taken over all centrally symmetric convex bodies \(K\) in \(\mathbb{R}^ d\). These lower bounds follow from the Minkowski inequality for mixed volumes, the Petty projection inequality, the Busemann intersection inequality and the Mahler-Reisner inequality. For 2-dimensional Minkowski spaces the lower bounds obtained are surprisingly close to the actual bound 6. For \(d=3\), and assuming Mahler's conjecture (i.e. \(m_ d={4^ d\over d }\)), the given lower bounds improve previous results. The exact values of \(\mu^{(1)}_ B(\partial B)\) and \(\mu^{(2)}_ B(\partial B)\) are calculated for three balls in \(\mathbb{R}^ d\), an ellipsoid, a cube and a zonotope. This yields several conjectures for which balls \(\mu^{(i)}_ B(\partial B)\) is maximal or minimal. Using the same inequalities the author further obtains a characterization of ellipsoids as those balls that contain their isoperimetrix.

Recommendations

Cites work

Cited in

(10)

This page was built for publication: Applications of various inequalities to Minkowski geometry

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1801635)