Metric spaces admitting low-distortion embeddings into all $n$-dimensional Banach spaces

Abstract

For a fixed $K\gg 1$ and $n\in \mathbb{N}$, $n\gg 1$, we study metric spaces which admit embeddings with distortion $\le K$ into each $n$-dimensional Banach space. Classical examples include spaces embeddable into $\log n$-dimensional Euclidean spaces, and equilateral spaces. We prove that good embeddability properties are preserved under the operation of metric composition of metric spaces. In particular, we prove that any $n$-point ultrametric can be embedded with uniformly bounded distortion into any Banach space of dimension $\log n$. The main result of the paper is a new example of a family of finite metric spaces which are not metric compositions of classical examples and which do embed with uniformly bounded distortion into any Banach space of dimension $n$. This partially answers a question of G. Schechtman.

37 pages, 5 figures, some small improvements of presentation

References in corpus (1)

• Embeddability of snowflaked metrics with applications to the nonlinear geometry of the spaces $L_p$ and ${\ell }_p$ for $0<p<\infty$