[Paper Review] On the easiest way to connect $k$ points in the Random Interlacements process

Johan Tykesson, Hubert Lacoin|arXiv (Cornell University)|Jun 19, 2012

Stochastic processes and statistical mechanics13 references20 citations

TL;DR

This paper determines the minimal number of trajectories required to almost surely connect any k ≥ 3 points in the random interlacement process on ℤ^d (d ≥ 5). It proves that n(k,d) = ⌈d/2(k−1)⌉ − (k−2) trajectories are sufficient and necessary in the worst case, establishing a sharp threshold using diagrammatic sums and path connectivity arguments in the interlacement model's trace structure.

ABSTRACT

We consider the random interlacements process with intensity $u$ on ${\\mathbb Z}^d$, $d\\ge 5$ (call it $I^u$), built from a Poisson point process on the space of doubly infinite nearest neighbor trajectories on ${\\mathbb Z}^d$. For $k\\ge 3$ we want to determine the minimal number of trajectories from the point process that is needed to link together $k$ points in $\\mathcal I^u$. Let $$n(k,d):=\\lceil \\frac d 2 (k-1) \ ceil - (k-2).$$ We prove that almost surely given any $k$ points $x_1,...,x_k\\in \\mathcal I^u$, there is a sequence ofof $n(k,d)$ trajectories $\\gamma^1,...,\\gamma^{n(k,d)}$ from the underlying Poisson point process such that the union of their traces $\\bigcup_{i=1}^{n(k,d)}\ r(\\gamma^{i})$ is a connected set containing $x_1,...,x_k$. Moreover we show that this result is sharp, i.e. that a.s. one can find $x_1,...,x_k in I^u$ that cannot be linked together by $n(k,d)-1$ trajectories.

Motivation & Objective

• To determine the minimal number of trajectories from the Poisson point process required to connect any k ≥ 3 points almost surely in the random interlacement set.
• To establish a sharp threshold for connectivity, showing that n(k,d) trajectories are both sufficient and necessary in the worst-case configuration.
• To extend prior results on two-point connectivity (k=2) to general k ≥ 3, completing the picture of finite-point connectivity in the model.
• To develop a novel diagrammatic sum approach for proving the lower bound, distinguishing it from prior capacity-based methods used in the k=2 case.
• To rigorously characterize the structure of trajectory traces that minimize connectivity cost, using tree-reduction techniques in the diagrammatic framework.

Proposed method

• Uses the random interlacement process constructed from a Poisson point process on doubly infinite nearest-neighbor trajectories on ℤ^d.
• Defines the trace of a trajectory as its image in ℤ^d, and considers connectivity of traces across multiple trajectories.
• Applies diagrammatic sum techniques to bound the probability of connecting k points using fewer than n(k,d) trajectories, proving the lower bound.
• Employs a tree-reduction procedure to simplify connectivity diagrams, reducing edge lengths and preserving critical decay properties in the sums.
• Uses the inequality (x−y)^{l−d}(z−y)^{l′−d} ≤ (x−y)^{l+l′−2d} + (z−y)^{l+l′−2d} to decompose complex interactions into simpler sub-diagrams.
• Relies on the fact that ∑_y (|x−y|+1)^{l−d}(|y−z|+1)^{l′−d} = O((|x−z|+1)^{l+l′+δ−d}) when l+l′ < d, enabling convergence of the sums.

Experimental results

Research questions

• RQ1 What is the minimal number of trajectories needed to almost surely connect any k ≥ 3 points in the random interlacement process on ℤ^d for d ≥ 5?
• RQ2 Is the proposed number of trajectories, n(k,d) = ⌈d/2(k−1)⌉ − (k−2), sharp in the sense that k points cannot be connected using n(k,d)−1 trajectories in some configurations?
• RQ3 How does the connectivity threshold for k points generalize the known two-point result (k=2) involving at most ⌈d/2⌉ trajectories?
• RQ4 Can diagrammatic sum techniques be used to prove a lower bound on the number of trajectories required for k-point connectivity?
• RQ5 What structural properties of trajectory traces and their interactions govern the minimal connectivity cost in the random interlacement model?

Key findings

• For any k ≥ 3 and d ≥ 5, almost surely, any k points in the random interlacement set can be connected via the traces of exactly n(k,d) = ⌈d/2(k−1)⌉ − (k−2) trajectories.
• The bound is sharp: there exist configurations of k points in the interlacement set that cannot be connected using only n(k,d)−1 trajectories.
• The result generalizes the k=2 case, where at most ⌈d/2⌉ trajectories are needed, to arbitrary finite k.
• The proof of the upper bound extends techniques from [RS10], while the lower bound relies on a novel diagrammatic sum argument with tree reduction.
• The diagrammatic method ensures convergence of the sums by controlling decay rates via edge length conditions and δ-regularization.
• The tree-reduction process preserves the total length of the diagram up to an arbitrarily small ε, enabling inductive control of the sum bounds.