English

Fluctuations of the connectivity threshold and largest nearest-neighbour link

Probability 2024-07-18 v3

Abstract

Consider a random uniform sample of nn points in a compact region AA of Euclidean dd-space, d2d \geq 2, with a smooth or (when d=2d=2) polygonal boundary. Fix kNk \in {\bf N}. Let Tn,kT_{n,k} be the threshold rr at which the geometric graph on these nn vertices with distance parameter rr becomes kk-connected. We show that if d=2d=2 then n(π/A)Tn,12lognn (\pi/|A|) T_{n,1}^2 - \log n is asymptotically standard Gumbel. For (d,k)(2,1)(d,k) \neq (2,1), it is n(θd/A)Tn,kd(22/d)logn(42k2/d)loglognn (\theta_d/|A|) T_{n,k}^d - (2-2/d) \log n - (4-2k-2/d) \log \log n that converges in distribution to a nondegenerate limit, where θd\theta_d is the volume of the unit ball. The limit is Gumbel with scale parameter 2 except when (d,k)=(2,2)(d,k)=(2,2) where the limit is two component extreme value distributed. The different cases reflect the fact that boundary effects are more more important in some cases than others. We also give similar results for the largest kk-nearest neighbour link Un,kU_{n,k} in the sample, and show Tn,k=Un,kT_{n,k}=U_{n,k} with high probability. We provide estimates on rates of convergence and give similar results for Poisson samples in AA. Finally, we give similar results even for non-uniform samples, with a less explicit sequence of centring constants.

Keywords

Cite

@article{arxiv.2406.00647,
  title  = {Fluctuations of the connectivity threshold and largest nearest-neighbour link},
  author = {Mathew D. Penrose and Xiaochuan Yang},
  journal= {arXiv preprint arXiv:2406.00647},
  year   = {2024}
}

Comments

39 pages, 3 figures. Various minor corrections and clarifications of v1