Curvilinear extensions of the Brunn-Minkowski-Lusternik inequality (Q1342764): Difference between revisions

The author gives a series of far reaching generalizations of the Brunn (-Minkowski-Lusternik) inequality for measurable sets in \(\mathbb{R}^{n + 1}\) in which instead of Minkowski linear combination a more general linear combination is used \[ \bigl( \lambda (x_i) {\overset \beta +} (1 - \lambda) (y_i) \bigr) = \bigl( \lambda x^\beta_i + (1 - \lambda) y^\beta_i \bigr)^{1/ \beta} \text{ for } (x_i), (y_i) \in \mathbb{R}^{n + 1} \text{ and } - \infty \leq \beta \leq \infty. \] In addition, an extension of Blaschke's ``Schüttelung'' is employed. Most of the results are too complicated to state in a review but the following theorem should give the reader an idea of the flavour of the results. For measurable \(A \subset \mathbb{R}^n \oplus [0, + \infty)\) let \(A_0 (\xi) = \{x \in \mathbb{R}^n : \mu (A \cap L(x)) \geq \xi \cdot \sup \{\mu (A \cap L(y)) : y \in \mathbb{R}^n\}\), \(A \cap L(x) \neq \emptyset\}\) and \(A_0 = \{(x, \eta) \in \mathbb{R}^n \oplus [0,+ \infty) : 0 \leq \eta \leq \mu (A \cap L(x))/ \sup\{ \{\mu (A \cap L(y)) : y \in \mathbb{R}^n\}\), \(A \cap L(x) \neq \emptyset\}\) where \(\mu\) is 1-dimensional Lebesgue measure and \(L(x) = \{(x, \eta) : \eta \geq 0\}\). Theorem 3.5: Let \(A,B \subset \mathbb{R}^n \oplus [0, + \infty)\) be bounded \(F_\sigma\)-sets of positive measure and let \(0 < \lambda < 1\). Then for \(- \infty \leq \alpha \leq + \infty\) the following inequalities hold \(\mu (\lambda A_0 {\overset \alpha +} (1 - \lambda) B_0) \geq \mu (\lambda A_0 {\overset - \infty +} (1 - \lambda) B_0) \geq \int^1_0 \mu (\lambda A_0 (\xi) + (1 - \lambda) B_0 (\xi)) d \xi\).