Affine quermassintegrals of random polytopes (Q2320093)

The affine quermassintegrals of a convex body \(K\) in \(\mathbb{R}^n\) are defined by \(\Phi_{n-k}(K) = \) \(\frac{\omega_n}{\omega_k}\left(\int_{G_{n,k}} vol_k(P_F(K))^{-n} d\nu_{n, k}(F)\right)^{-1/n}\) for \(1\le k \le n-1\), where \(\nu_{n, k}\) is the Haar probability measure on the Grassmannian \(G_{n,k}\) of all \(k\)-dimensional subspaces of \(\mathbf{R}^n\) and \(\omega_k\) is the volume of the Euclidean unit ball \(B_2^k\) and \(\Phi_0(K) = vol_n(K)\), \(\Phi_n(K) = \omega_n\). \textit{N. Dafnis} and \textit{G. Paouris} [Ill. J. Math. 56, No. 4, 1005--1021 (2012; Zbl 1335.52008)] studied an isomorphic variant of Lutwak's conjecture. They ask if there exist absolute positive constants \(c_1\), \(c_2\) such that for every convex body \(K\) and any \(1\le k\le n-1\) we have \(c_1 \sqrt{n/k} \le \Phi_{[k]}(K) \le c_2 \sqrt{n/k}\). The main question discussed in the present paper is related to the upper bound in the above inequality. An almost optimal estimate (up to a \(\log\) \(n\)-term) was given in [\textit{S. Artstein-Avidan} et al., Asymptotic geometric analysis. I. Providence, RI: American Mathematical Society (AMS) (2015; Zbl 1337.52001)]. It was shown there that \(\Phi_{[k]}(K)\le c_3 (\sqrt{n/k})^3 \sqrt{\log (e n/k)}\). In other words, if \(k\) is proportional to \(n\) then the upper bound for \(\Phi_{[k]}(K)\) is of order of 1. The main question that remains open is whether the \(\log n\)-term in the previous inequality can actually be dropped. In the present paper the authors investigate this question for some broad classes of random polytopes. They provide an affirmative answer to the problem for the class of symmetric random polytopes with at most \(e^{\sqrt{n}}\) vertices uniformly distributed in a convex body.