Random aspects of high-dimensional convex bodies (Q2702148)

The authors study geometry of compact, not necessarily centrally symmetric, convex bodies in \(R^n\). They concentrate on random aspects of compact convex bodies and investigate some invariants fundamental in the local theory of Banach spaces, restricted to random sections and projections of such bodies. It turns out that, loosely speaking, such random operations kill the effect of non-symmetry in the sense that resulting estimates are very close to their centrally symmetric counterparts. First, the authors develop some tools. They introduce the concept of random Gelfand numbers, which formalizes the phenomenon of lower estimates by the Euclidean norm on random subspaces. They prove that for any convex body \(K\subset R^n\), good lower estimates on subspaces of proportional dimensions imply similar estimates valid on random subspaces of comparable dimensions. Then the authors prove that every convex body \(K\) has a position \(K_1\) (i.e., \(K_1=uK-a\) for some operator \(u\) and \(a\in R^n\)) such that a random projection \(PK_1\), of dimension \([n/2]\) satisfies \(M(PK_1)M(K^0_1)\leq C\log n\), where \(C\) is an absolute constant. Moreover, there exists a unitary operator \(u\) such that \(M(K_1+uK_1)M(K_1^0)\leq C\log n\). The proof is based essentially on symmetric considerations, a non-symmetric part is reduced to classical facts and simple lemmas. Next, they consider a natural way to measure the ``non-symmetry'' of a convex body, as the distance of a body to the set of all centrally symmetric bodies. It turns out that for random projections of rank \(k\) of a simplex \(S\subset R^n\), this distance can be asymptotically estimated from below by \(\sqrt{k/\ln n}\). On the other hand, the case of simplex is the worst. Finally, the authors discuss the so-called proportional Dvoretzky-Rogers factorization in the non-symmetric setting. The proof essentially follows the symmetric argument, but has a few delicate points and it needs some extra work. Some applications of this are shown at the end of the paper. For the entire collection see [Zbl 0949.00025].