English

Hyperuniform and rigid stable matchings

Probability 2020-01-29 v3 Disordered Systems and Neural Networks

Abstract

We study a stable partial matching τ\tau of the (possibly randomized) dd-dimensional lattice with a stationary determinantal point process Ψ\Psi on Rd\mathbb{R}^d with intensity α>1\alpha>1. For instance, Ψ\Psi might be a Poisson process. The matched points from Ψ\Psi form a stationary and ergodic (under lattice shifts) point process Ψτ\Psi^\tau with intensity 11 that very much resembles Ψ\Psi for α\alpha close to 11. On the other hand Ψτ\Psi^\tau is hyperuniform and number rigid, quite in contrast to a Poisson process. We deduce these properties by proving more general results for a stationary point process Ψ\Psi, whose so-called matching flower (a stopping set determining the matching partner of a lattice point) has a certain subexponential tail behaviour. For hyperuniformity, we also additionally need to assume some mixing condition on Ψ\Psi. Further, if Ψ\Psi is a Poisson process then Ψτ\Psi^\tau has an exponentially decreasing truncated pair correlation function.

Keywords

Cite

@article{arxiv.1810.00265,
  title  = {Hyperuniform and rigid stable matchings},
  author = {Michael Andreas Klatt and Günter Last and D. Yogeshwaran},
  journal= {arXiv preprint arXiv:1810.00265},
  year   = {2020}
}

Comments

Pages: 35, Figures: 10, Supplementary Video: Zooming out of a Poisson point process and a hyperuniform and rigid stable matching, Supplementary Animations: Visualization of the mutual nearest neighbor matching algorithm and the construction of a matching flower, Second version: Some typos and minor inaccuracies have been corrected

R2 v1 2026-06-23T04:23:09.796Z