Supercritical phase of the random connection model
Penrose · math.PR · 2026-09-05 · 原文
Given d \in {\bf N}, λ>0, the random connection model (RCM) in a region A \subseteq {\bf R}^d is a graph with vertex set given by a homogeneous Poisson point process of intensity λ in A, with an edge placed between each pair x,y of vertices with probability ϕ(\|x-y\|), where ϕ: {\bf R}_+ \to [0,1] is a nonincreasing finite-range connection function. We show that if d \geq 3 and λ is strictly supercritical for the RCM in {\bf R}^d, then the model remains supercritical if it is restricted to a region A of the form {\bf R}^2 \times [-K/2,K/2]^{d-2}, provided K is sufficiently large. This is a continuum analogue of a well-known result of Grimmett and Marstrand for lattice percolation. We prove this by adapting Grimmett and Marstrand's original proof; Faggionato and Hartarsky have also proved this recently by other means.
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
Given d \in {\bf N}, λ>0, the random connection model (RCM) in a region A \subseteq {\bf R}^d is a graph with vertex set given by a homogeneous Poisson point process of intensity λ in A, with an edge placed between each pair x,y of vertices with probability ϕ(\|x-y\|), where ϕ: {\bf R}_+ \to [0,1] is a nonincreasing finite-range connection function.
We show that if d \geq 3 and λ is strictly supercritical for the RCM in {\bf R}^d, then the model remains supercritical if it is restricted to a region A of the form {\bf R}^2 \times [-K/2,K/2]^{d-2}, provided K is sufficiently large.
2. 领域脉络
本文类目：math.PR，属于其所在研究脉络的最新进展。
3. 机制拆解
This is a continuum analogue of a well-known result of Grimmett and Marstrand for lattice percolation.
We prove this by adapting Grimmett and Marstrand's original proof; Faggionato and Hartarsky have also proved this recently by other means.
4. 证据与数字
摘要未给出量化结果——留意原文的实验与数据。
5. 反例与边界
摘要未声明局限与反例——这是需要警惕的信号，精读时先问边界。
6. 跨领域连接与意外收获
思考本文机制能否迁移到你正在跟进的问题。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。