The isotropic position and the reverse Santaló inequality (Q476488): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Set profile property. |
||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank |
Revision as of 00:24, 5 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | The isotropic position and the reverse Santaló inequality |
scientific article |
Statements
The isotropic position and the reverse Santaló inequality (English)
0 references
2 December 2014
0 references
The classical Blaschke-Santaló inequality states that for every symmetric convex body \(K\) in \(\mathbb{R}^n\) the volume product \(s(K)=|K||K^{\circ}|\) is bounded above by \(s(B_2^n)\), with equality if and only if \(K\) is an ellipsoid. Here \(|\cdot|\) denotes the volume and \(B_2^n\) the unit Euclidean ball. The reverse Santaló inequality, first proven in [\textit{J. Bourgain} and \textit{V. D. Milman}, Invent. Math. 88, 319--340 (1987; Zbl 0617.52006)], establishes the existence of an absolute constant \(c>0\) such that for any convex body \(K\), whose interior contains \(0\), the following inequality holds \[ \left(\frac{s(K)}{s(B_2^n)}\right)^{1/n}\geq c. \] There are several proofs of this inequality which use tools from very different areas. In this beautiful paper, the authors provide a new approach to the reverse Santaló inequality based on properties of the isotropic position of a convex body. This approach combines an isomorphic symmetrization argument, similar to the one used by \textit{V. D. Milman} in [Lect. Notes Math. 1317, 107--131 (1988; Zbl 0652.52008)], with the \textit{method of convex perturbations}, introduced by \textit{B. Klartag} in [Geom. Funct. Anal. 16, No. 6, 1274--1290 (2006; Zbl 1113.52014)]. The approach uses only purely convex geometrical tools. The new ingredient (according to the authors) is a way to show that every convex body with bounded isotropic constant satisfies the reverse Santaló inequality, and then that all bodies do. The proof is made in two major steps. First, a lower bound for the volume product \(s(K)\) is obtained which involves the isotropic constant \(L_K\); it is done using covering estimates: Let \(K\) be a convex body in \(\mathbb{R}^n\) which contains \(0\) in its interior. Then \[ 4n s(K)^{1/n}\geq n s(K-K)^{1/n}\geq \frac{c_1}{L_K}, \] where \(c_1>0\) is an absolute constant. Next, the authors manage to adapt the main argument in the above mentioned work by Klartag to show that \(K\) is geometrically close to a convex body \(T\), whose isotropic constant is related to \(s(K)\): Let \(K\) be a symmetric convex body in \(\mathbb{R}^n\). There exists a convex body \(T\subset \mathbb{R}^n\) such that \[ c_2 K\subseteq T-T\subseteq c_3 K \;\text{and}\; L_T\leq\frac{c_4}{\sqrt{n s(K)^{1/n}}}, \] where \(c_2,c_3,c_4\) are absolute constants. Joining this with the bound for \(L_T\), the isotropic constant can be removed from the bound and the reverse Santaló inequality is proved: Let \(K\) be a symmetric convex body in \(\mathbb{R}^n\). Then \(s(K)^{1/n}\geq c_5/n\), where \(c_5\) is an absolute constant. Using this, in the last section the authors establish the existence of \(M\)-ellipsoids. As a consequence, the reverse Brunn-Minkowski inequality is proved. The authors make a brief nice historical tour about the results and provide details for all required tools from asymptotic convex geometry.
0 references
reverse Santaló inequality
0 references
isotropic position
0 references
\(M\)-ellipsoid
0 references
reverse Brunn-Minkowski inequality
0 references
log-concave measure
0 references
covering number
0 references
convex perturbation
0 references