The floating body in real space forms (Q1799685): Difference between revisions

Let \(\mathrm{Sp}^{n}(\lambda)\) denote a simply connected complete Riemannian manifold with constant sectional curvature \(\lambda\) (i.e., a real space form). A compact and geodesically convex set \(K\subset\mathrm{Sp}^{n}(\lambda)\) is called a convex body, and the set of convex bodies with nonempty interior is denoted by \(\mathcal{K}_0(\mathrm{Sp}^{n}(\lambda))\). For \(\delta >0\), the \(\lambda\)-floating body \(\mathcal{F}_{\delta}^{\lambda}K\) of \(K\in\mathcal{K}_0(\mathrm{Sp}^{n}(\lambda))\) is defined as the intersection of all half-spaces \(H^-\) cutting off a set of volume less or equal \(\delta\) from \(K\), i.e., \[ \mathcal{F}_{\delta}^{\lambda}K=\bigcap\left\{H^-\,:\,\mathrm{vol}_n^{\lambda}(K\cap H^+)\leq \delta^{\frac{n+1}{2}}\right\}, \] where \(H\) denotes a hyperplane in \(\mathrm{Sp}^{n}(\lambda)\) which divides the space into two half-spaces \(H^+\) and \(H^-\). The authors study the volume of the floating body of \(K\in\mathcal{K}_0(\mathrm{Sp}^{n}(\lambda))\). More precisely, they prove that the right-derivative of \(\mathrm{vol}_n^{\lambda}(F_{\delta}^{\lambda}K)\) at \(\delta=0\) exists and equals \(\Omega^{\lambda}(K)\), called the \(\lambda\)-floating area of \(K\), in the paper under review. For \(\lambda=0\), this coincides with the affine surface area, and the result of the paper was first proved in [\textit{C. Schütt} and the second author, Math. Scand. 66, No. 2, 275--290 (1990; Zbl 0739.52008)]. For \(\lambda=1\), the result has also been proved in [the authors, Adv. Math. 301, 867--901 (2016; Zbl 1353.52006)]. The \(\lambda\)-floating area \(\Omega^{\lambda}:\mathcal{K}_0(\mathrm{Sp}^{n}(\lambda))\to\mathbb{R}\) is shown to be upper semi-continuous, a valuation, and invariant under the isometries of \(\mathrm{Sp}^{n}(\lambda)\). In the last section, the authors ask for a generalization of the affine isoperimetric inequality in \(\mathrm{Sp}^{n}(\lambda)\) by using \(\Omega^{\lambda}\) and conjecture that the optimizers are, for all \(\lambda\), the geodesic balls.