On rough isometries of Poisson processes on the line

  • Ron Peled*
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

8 Scopus citations

Abstract

Intuitively, two metric spaces are rough isometric (or quasi-isometric) if their large-scale metric structure is the same, ignoring fine details. This concept has proven fundamental in the geometric study of groups. Abért, and later Szegedy and Benjamini, have posed several probabilistic questions concerning this concept. In this article, we consider one of the simplest of these: are two independent Poisson point processes on the line rough isometric almost surely? Szegedy conjectured that the answer is positive. Benjamini proposed to consider a quantitative version which roughly states the following: given two independent percolations on ℕ, for which constants are the first n points of the first percolation rough isometric to an initial segment of the second, with the first point mapping to the first point and with probability uniformly bounded from below? We prove that the original question is equivalent to proving that absolute constants are possible in this quantitative version. We then make some progress toward the conjecture by showing that constants of order √ log n suffice in the quantitative version. This is the first result to improve upon the trivial construction which has constants of order log n. Furthermore, the rough isometry we construct is (weakly) monotone and we include a discussion of monotone rough isometries, their properties and an interesting lattice structure inherent in them.

Original languageEnglish
Pages (from-to)462-494
Number of pages33
JournalAnnals of Applied Probability
Volume20
Issue number2
DOIs
StatePublished - Apr 2010
Externally publishedYes

Keywords

  • Geometry of random sets
  • Matching
  • One dimension
  • Percolation
  • Poisson process
  • Quasi-isometry
  • Rough isometry

Fingerprint

Dive into the research topics of 'On rough isometries of Poisson processes on the line'. Together they form a unique fingerprint.

Cite this