[Paper Review] The Gaussian free field in interlacing particle systems

Jeffrey Kuan|arXiv (Cornell University)|Sep 20, 2011

Stochastic processes and statistical mechanics14 references4 citations

TL;DR

This paper establishes that fluctuations in a specific interlacing particle system with a reflecting wall converge to the Gaussian free field (GFF) under broad conditions: when the system is a determinantal point process with a correlation kernel expressible as a double integral. The key result confirms that a previously studied random surface growth model exhibits GFF fluctuations, linking it to universal statistical behavior in integrable probability and statistical mechanics.

ABSTRACT

We show that if an interlacing particle system in a two-dimensional lattice is a determinantal point process, and the correlation kernel can be expressed as a double integral with certain technical assumptions, then the moments of the fluctuations of the height function converge to that of the Gaussian free field. In particular, this shows that a previously studied random surface growth model with a reflecting wall has Gaussian free field fluctuations.

Motivation & Objective

• To determine whether fluctuations in a specific interlacing particle system with a reflecting wall converge to the Gaussian free field.
• To establish general conditions under which the moments of height function fluctuations in such systems converge to those of the GFF.
• To confirm that a known random surface growth model with a wall has universal GFF fluctuations, linking it to broader universality principles in statistical mechanics.
• To provide a rigorous asymptotic analysis of the correlation kernel via complex analysis and Laplace method techniques.

Proposed method

• The analysis relies on the system being a determinantal point process with a correlation kernel expressible as a double integral over contours in the complex plane.
• The height function is defined as the number of particles to the right of a given point, and its fluctuations are studied in the large-N limit.
• The proof uses the method of steepest descent and complex analysis to evaluate asymptotic integrals of the kernel, focusing on critical points of a generating function G(z).
• Key estimates are derived using bounds on derivatives of the function R(t), which captures the logarithmic rate of the kernel's behavior near its maximum.
• Watson’s lemma and Laplace method approximations are applied to control error terms in the asymptotic expansion of integrals involving the kernel.
• The convergence of moments is established by comparing the moments of the height function to those of the GFF through bounds on the kernel’s behavior near critical points.

Experimental results

Research questions

• RQ1 Under what conditions does the height function of an interlacing particle system converge to the Gaussian free field in the large-N limit?
• RQ2 Can the GFF fluctuations be rigorously established for a random surface growth model with a reflecting wall?
• RQ3 How do the moments of the height function fluctuations behave asymptotically when the correlation kernel admits a double integral representation?
• RQ4 What role does the critical point of the function G(z) play in determining the limit shape and fluctuation behavior?
• RQ5 Can the error terms in the asymptotic expansion of the kernel be controlled to ensure convergence to the GFF?

Key findings

• The moments of the height function fluctuations converge to those of the Gaussian free field under the stated technical conditions on the kernel and the domain of analyticity.
• The limit shape of the system is explicitly described via a critical point map Ω(ν,η,τ) into the upper half-plane minus the unit disk.
• The fluctuations are universal in the sense that they match the GFF regardless of the specific details of the particle dynamics, provided the kernel satisfies the double integral condition.
• The error terms in the asymptotic expansion are bounded by O(N^{-3/2}) under the given assumptions, ensuring convergence of moments.
• The analysis confirms that the random surface growth model with a reflecting wall, previously studied in the literature, exhibits GFF fluctuations.
• The result extends to the edge of the liquid region, with improved error bounds obtained via a refined application of Watson’s lemma.