Support functions and mean width for \(\alpha\)-concave functions (Q392959)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Support functions and mean width for \(\alpha\)-concave functions |
scientific article |
Statements
Support functions and mean width for \(\alpha\)-concave functions (English)
0 references
15 January 2014
0 references
Denote by \(\mathrm{Cvx}({\mathbb{R}}^n)\) the set of all convex and lower semicontinuous functions \(\phi : {\mathbb{R}}^n \to (-\infty, \infty]\) with \(\phi (x)<\infty \) for an \(x\in {\mathbb{R}}^n\). One of the main results of this paper is the following. {Theorem.} Assume that a function \(T: \mathrm{Cvx}({\mathbb{R}}^n)\to \mathrm{Cvx}({\mathbb{R}}^n)\) satisfies the following three conditions: (1) \(T\) is order preserving: \(\phi\leq \psi\) if and only if \(T \phi \leq T \psi\), (2) if \(\phi\) is a positively homogeneous function, then \(T \phi = \phi\), and (3) the set \(\{T \phi : \phi \in \mathrm{Cvx}({\mathbb{R}}^n)\}\) is closed under pointwise addition. Then \((T \phi)(x) = \frac{1}{C}\phi (Cx)\) for a positive \(C\). From this theorem a characterization of the support function and the sup-convolution operation is obtained. Moreover, some notions are extended from log-concave functions to \(\alpha\)-concave functions. For non-zero, \(\alpha\)-concave and upper semicontinuous functions mapping \({\mathbb{R}}^n\) into \([0, \infty)\) the author defines \(\alpha\)-support function, \(\alpha\)-sum, \(\lambda\)-homothety and \(\alpha\)-mean width, and gives a few of their properties. One of them is an Urysohn type inequality for \(\alpha\)-mean width, which is applied to get a generalized Poincaré inequality.
0 references
\(\alpha\)-concavity
0 references
mean width
0 references
support function
0 references
sup-convolution operation
0 references
Urysohn type inequality
0 references
generalized Poincaré inequality
0 references
0 references