English

Localization in random geometric graphs with too many edges

Probability 2019-07-04 v7

Abstract

We consider a random geometric graph G(χn,rn)G(\chi_n, r_n), given by connecting two vertices of a Poisson point process χn\chi_n of intensity nn on the unit torus whenever their distance is smaller than the parameter rnr_n. The model is conditioned on the rare event that the number of edges observed, E|E|, is greater than (1+δ)E(E)(1 + \delta)\mathbb{E}(|E|), for some fixed δ>0\delta >0. This article proves that upon conditioning, with high probability there exists a ball of diameter rnr_n which contains a clique of at least 2δE(E)(1ε)\sqrt{2 \delta \mathbb{E}(|E|)}(1 - \varepsilon) vertices, for any given ε>0\varepsilon >0. Intuitively, this region contains all the "excess" edges the graph is forced to contain by the conditioning event, up to lower order corrections. As a consequence of this result, we prove a large deviations principle for the upper tail of the edge count of the random geometric graph. The rate function of this large deviation principle turns out to be non-convex.

Keywords

Cite

@article{arxiv.1401.7577,
  title  = {Localization in random geometric graphs with too many edges},
  author = {Sourav Chatterjee and Matan Harel},
  journal= {arXiv preprint arXiv:1401.7577},
  year   = {2019}
}

Comments

56 pages, 1 figure

R2 v1 2026-06-22T02:57:11.746Z