Generalized transportation cost spaces

From MaRDI portal
Publication:2295447



Abstract: The paper is devoted to the geometry of transportation cost spaces and their generalizations introduced by Melleray, Petrov, and Vershik (2008). Transportation cost spaces are also known as Arens-Eells, Lipschitz-free, or Wasserstein 1 spaces. In this work, the existence of metric spaces with the following properties is proved: (1) uniformly discrete infinite metric spaces transportation cost spaces on which do not contain isometric copies of ell1, this result answers a question raised by Cuth and Johanis (2017); (2) locally finite metric spaces which admit isometric embeddings only into Banach spaces containing isometric copies of ell1; (3) metric spaces for which the double-point norm is not a norm. In addition, it is proved that the double-point norm spaces corresponding to trees are close to ellinftyd of the corresponding dimension, and that for all finite metric spaces M, except a very special class, the infimum of all seminorms for which the embedding of M into the corresponding seminormed space is isometric, is not a seminorm.


The authors present two elementary constructions of metric spaces. \begin{enumerate} \item An infinite uniformly discrete metric space \(M_1\) such that the transportation cost space (a.k.a. free space, Arens-Eells space, \dots) over \(M_1\) does not contain an isometric copy of \(\ell_1\). The metric space \(M_1\) is the set \(\mathbb{N}\) equipped with the metric \(d_h(i,j)=h(\min\{i,j\})\) if \(i\neq j\) and \(d_h(i,j)=0\) if \(i=j\), where \(h: \mathbb{N}\to (1,2)\) is a strictly increasing function. \item An infinite locally finite metric space \(M_2\) such that every Banach space that contains the \(\ell_1^n\)'s isometrically but not \(\ell_1\) isometrically, will contain an isometric copy of all finite subsets of \(M_2\) but not an isometric copy of \(M_2\). In fact, it is shown that, if \(M_2\) embeds isometrically into an infinite-dimensional Banach space \(Y\), then \(Y\) must necessarily contain an isometric copy of \(\ell_1\). The metric space \(M_2\) is the subset of \(\ell_1\) consisting of vectors of the form \(\sum_{i\in A}2^ie_i\), where \(A\) is a finite subset of \(\mathbb{N}\), and \((e_i)_{i\ge 1}\) denotes the canonical basis of \(\ell_1\). \end{enumerate} The rest of the paper is a technical investigation of various semi-norms on the vector space of transportation plans on a metric space \(M\), i.e., the vector space of all finitely supported functions \(f: M\to \mathbb{R}\) such that \(\sum_{x\in M}f(x)=0\).



Cites work


Cited in
(21)


Describes a project that uses

Uses Software






This page was built for publication: Generalized transportation cost spaces

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2295447)