Long-range contact process and percolation on a random lattice

Basic Information

• Paper ID: 2108.13531
• Title: Long-range contact process and percolation on a random lattice
• Authors: Pablo A. Gomes (Universidade de São Paulo), Bernardo N. B. de Lima (Universidade Federal de Minas Gerais)
• Classification: math.PR (Probability Theory)
• Publication Date: July 20, 2022 (arXiv v3)
• Paper Link: https://arxiv.org/abs/2108.13531

Abstract

This paper investigates phase transition phenomena in long-range oriented percolation and contact processes. The authors study a contact process where the infection range of each vertex is independent, dynamically updated, and governed by some distribution N. Simultaneously, they investigate an analogous oriented percolation model on hypercubic lattices, where a special direction permits long-range oriented edges; the ranges of all vertices are given by independent identically distributed random variables with common distribution N. For both models, the authors establish results concerning the existence of phase transitions, which depend on the properties of distribution N.

Research Background and Motivation

Importance of the Problem

1. Physical significance of long-range models: Long-range models hold important positions in mathematical physics literature. Prior to percolation theory, Dyson and Frolich-Spencer studied long-range Ising models. Such models better describe real-world interacting systems.
2. Theoretical value of phase transitions: Phase transitions are central concepts in statistical physics and probability theory. Understanding phase transition behavior under long-range interactions is crucial for theoretical development.
3. Social network applications: Long-range percolation models have proven to be effective tools for studying social networks, particularly in investigating graph distances (chemical distances) on long-range percolation clusters.

Limitations of Existing Research

1. Static range restrictions: Traditional long-range contact processes typically consider fixed infection ranges, failing to capture dynamically changing ranges in reality.
2. Deterministic parameters: Existing models predominantly use deterministic long-range connection probabilities, lacking systematic study of random ranges.
3. Dimensional limitations: Many results apply only to specific dimensions, lacking a unified theoretical framework.

Core Contributions

1. Introduction of dynamic-range contact process (CPDR): First systematic study of contact process models where infection ranges change dynamically over time.
2. Establishment of random-range percolation model (APRR): Proposes an anisotropic percolation model with random long-range connection ranges.
3. Complete characterization of phase transitions: For both models, provides complete phase transition criteria based on the range distribution N:
   • When EN^d < ∞, there exists a non-trivial critical infection rate
   • When limsup_{n→∞} nP(N^d ≥ n) > 0, positive survival probability exists for all infection rates
4. Unified proof techniques: Develops analytical methods applicable to long-range models in dynamic environments.

Methodology Details

Task Definition

CPDR Model Definition

Define a contact process on the complete graph G_d = (Z^d, E) where:

• λ > 0 is the infection rate parameter
• N is a range random variable taking values in Z_+
• The infection range r_x(t) of each vertex x at time t is dynamically updated by independent Poisson processes

APRR Model Definition

Define a percolation process on graph G = (Z^d, E_v ∪ E_h) where:

• E_h consists of long-range oriented edges along the first coordinate axis
• E_v consists of nearest-neighbor oriented edges in other directions
• The long-range connection range of each vertex is given by independent identically distributed random variables N_x

Model Architecture

Dynamic Mechanism of CPDR

1. Recovery process: Each vertex x has an independent Poisson recovery process R_x
2. Range update: Each vertex x has an independent Poisson update process T_x controlling range change timing
3. Infection propagation: At time t, infected vertex x can infect healthy vertices at distance not exceeding r_x(t)

The range update rule is:

r_x(t) = N_{x,n}, ∀t ∈ [S_{x,n-1}, S_{x,n})

where S_{x,n} = ∑^n T{x,k} is the n-th update time.

Geometric Structure of APRR

The random subgraph G_N is defined as:

G_N := (Z^d, E_v ∪ (∪_{x∈Z^d}{(x, x+n⃗e_1) ∈ Z^d × Z^d : n ≤ N_x}))

Technical Innovations

Atom Covering Technique

For proving the subcritical phase of CPDR, introduces the "atom" concept:

• Each atom A = (x, I) consists of vertex x and time interval I
• Atom sequences cover all possible infection paths
• Branching processes control atom generation

Block Renormalization Method

For proving the supercritical phase, employs innovative block renormalization technique:

• Partitions spacetime into disjoint blocks B_v = V_v × Δ_v
• Defines connection rules between blocks, establishing correspondence with percolation on renormalized lattices
• Controls connection probabilities to prove supercriticality of renormalized percolation

Experimental Setup

Theoretical Analysis Framework

This paper employs rigorous probabilistic proof methods, with main tools including:

1. Branching process theory: For controlling subcritical phase analysis
2. Percolation theory: Establishing connections with classical percolation models
3. Poisson process theory: Handling dynamic update mechanisms
4. Graph-theoretic methods: Analyzing geometric properties of long-range connections

Key Parameter Settings

• Infection rate λ as the primary control parameter
• Moment conditions of range distribution N as phase transition criteria
• Spatial dimension d affecting power exponents in critical conditions

Experimental Results

Main Theoretical Results

Phase Transition Theorem for CPDR Model

Theorem 1.1 (Subcritical phase): If EN^d < ∞, then there exists sufficiently small λ_0 such that for all 0 < λ < λ_0:

P(ζ_t ≠ ∅, ∀t ≥ 0) = 0

Theorem 1.2 (Supercritical phase): If limsup_{n→∞} nP(N^d ≥ n) > 0, then:

P(ζ_t ≠ ∅, ∀t ≥ 0) > 0, ∀λ > 0

Phase Transition Theorem for APRR Model

Theorem 1.3: If EN < ∞, then for all p < 1 we have q_c(p) > 0.

Theorem 1.4: If limsup_{n→∞} nP(N ≥ n) > 0, then for all p > 0 we have q_c(p) = 0.

Exact Results for Special Cases

For the distribution P(N ≥ n) = 1 - e^{-β/n}:

Theorem 1.5: When p = 1, q = 0, we have θ(1,0) = 0 for β ≤ 1 and θ(1,0) > 0 for β > 1.

Theorem 1.6: When 0 < p < 1, q = 0, we have θ(p,0) > 0 for β > p^{-1}.

Experimental Findings

1. Precise characterization of critical conditions: The tail behavior of range distribution completely determines the existence of phase transitions
2. Dimensional effects: Spatial dimension d operates through its influence on critical moment condition EN^d
3. Impact of dynamics: Dynamic update mechanisms do not alter the fundamental structure of phase transitions but affect critical parameters

Related Work

Historical Development of Long-Range Percolation

1. One-dimensional case: Schulman proved no percolation for s > 2; Newman-Schulman proved oriented percolation exists for s < 2
2. Critical case: Aizenman-Newman and Marchetti-Sidoravicius-Vares resolved the s = 2 critical case
3. High-dimensional generalization: Meester-Steif and others studied truncation issues in high-dimensional long-range percolation

Development of Contact Process Theory

1. Classical theory: Harris introduced the contact process in 1974 as an infection transmission model
2. Long-range variants: Spitzer and Bramson-Gray studied phase transitions in long-range contact processes
3. Dynamic environments: Recent work by Linker-Remenik and others studied contact processes in dynamic environments

Innovation of This Paper

This paper is the first to combine long-range interactions with dynamic random environments, providing a novel theoretical framework and analytical methods.

Conclusions and Discussion

Main Conclusions

1. Complete phase transition characterization: Provides complete phase transition criteria for both models based on tail behavior of range distribution
2. Unified theoretical framework: Develops general analytical methods applicable to dynamic long-range models
3. Precise critical phenomena: Provides exact critical parameters in special cases

Limitations

1. Dynamic requirements: The proof of Theorem 1.2 requires positive update rates, not applicable to static cases
2. Dimensional restrictions: Some results may require additional technical treatment for low-dimensional cases
3. Distribution assumptions: Results depend on specific tail behavior assumptions

Future Directions

1. Static case: Investigate phase transition behavior when update rate is zero
2. Geometric properties: Analyze geometric characteristics of long-range percolation clusters such as chemical distances
3. Application extensions: Apply results to practical network transmission models

In-Depth Evaluation

Strengths

1. Theoretical innovation: First systematic study of long-range models with dynamic ranges, filling theoretical gaps
2. Advanced methods: Develops new analytical techniques such as atom covering and block renormalization
3. Complete results: Provides complete phase transition characterization with nearly necessary and sufficient conditions
4. Rigorous proofs: Mathematical proofs are complete and rigorous with refined technical treatment

Weaknesses

1. Practical applications: The connection between theoretical models and practical application scenarios could be strengthened
2. Computational complexity: Some proof techniques are complex with room for improved readability
3. Numerical verification: Lacks numerical simulations to verify theoretical results

Impact

1. Theoretical contribution: Makes important contributions to the theory of long-range interaction systems
2. Method value: Developed analytical methods applicable to other related models
3. Application potential: Has potential applications in network science and epidemic modeling

Applicable Scenarios

1. Theoretical research: Applicable to theoretical research in probability theory and statistical physics
2. Network analysis: Can be used to analyze complex networks with dynamic connection ranges
3. Transmission modeling: Applicable to modeling transmission processes with varying influence ranges

References

This paper cites important literature in the field, including:

• Harris (1974): Pioneering work on contact processes
• Aizenman-Newman (1986): Classical results on one-dimensional long-range percolation
• Liggett (1985, 1999): Standard textbooks on interacting particle systems
• And numerous related works on long-range models and phase transition theory

Summary: This is a high-quality theoretical paper in probability theory that makes important contributions to phase transition theory in long-range interaction systems. The paper introduces new models, develops new analytical methods, and provides complete theoretical characterization. While primarily theoretical, it establishes a solid foundation for further development in related fields.