[Paper Review] The heat and the landscape I

Bálint Virág|arXiv (Cornell University)|Aug 17, 2020

Random Matrices and Applications19 references19 citations

TL;DR

This paper establishes that the O'Connell-Yor polymer and the KPZ equation in 1+1 dimensions converge to the KPZ fixed point under scaling, using a novel characterization of the fixed point via Baik-Ben Arous-Peche (BBP) statistics. The key contribution is a general, elementary method for proving convergence based on previously established limit theorems, leveraging the directed landscape and Airy sheet structure to show universal KPZ behavior across models.

ABSTRACT

Heat flows in 1+1 dimensional stochastic environment converge after scaling to the random geometry described by the directed landscape. In this first part, we show that the O'Connell-Yor polymer and the KPZ equation converge to the KPZ fixed point. The key is that one-dimensional Baik-Ben Arous-Peche statistics characterize the KPZ fixed point. This yields a general and elementary method that shows convergence based on previously established limit theorems. Independently, at the same time and place Quastel and Sharkar gave an unrelated proof of KPZ convergence. The methods invite extensions in different directions: ours to polymer models, and theirs to interacting particle systems.

Motivation & Objective

• To establish universal convergence of 1+1 dimensional stochastic models—specifically the O'Connell-Yor polymer and the KPZ equation—to the KPZ fixed point under scaling.
• To provide a new characterization of the KPZ fixed point using Baik-Ben Arous-Peche (BBP) statistics as a basis for convergence proofs.
• To develop a general and elementary method for proving convergence to the KPZ fixed point by leveraging existing limit theorems on edge statistics.
• To demonstrate that the directed landscape and Airy sheet structure underlie the universal scaling limit of diverse models, including polymers and SPDEs.
• To offer a framework that extends beyond the specific models, enabling future applications to other polymer models and stochastic systems.

Proposed method

• The paper uses a new characterization of the KPZ fixed point based on distance laws defined via metric composition with upper semicontinuous functions and the Airy sheet.
• It applies the functional form $ f \cdot \mathcal{S} \cdot g $, where $ \mathcal{S} $ is the Airy sheet, to define the fixed point's law through cumulative distribution functions.
• The convergence proof relies on showing that the law of $ f \cdot \mathcal{K}_n \cdot g $ converges to $ f \cdot \mathcal{S} \cdot g $ for test functions $ f, g $, using the tightness and metric Burke property of rescaled models.
• It establishes the metric Burke property $ \mathcal{B}_{\nu,n} \circ_n \mathcal{R} \stackrel{d}{=} \mathcal{R} \circ_n \mathcal{B}_{\nu,n} $ for rescaled Brownian motions and the stochastic kernel $ \mathcal{K}_n $, ensuring invariance under composition.
• The method uses the quadrangle inequality and compact convergence to control the behavior of the rescaled free energy and show tightness in space-time.
• It applies the 1-2-3 scaling and Cole-Hopf transformation to map the KPZ equation to the stochastic heat equation, enabling convergence to the fixed point via known limit theorems.

Experimental results

Research questions

• RQ1 Does the O'Connell-Yor polymer model converge to the KPZ fixed point under appropriate scaling, and if so, via what mechanism?
• RQ2 Can the KPZ fixed point be characterized solely by the laws of its distance distributions involving pointed edge processes and the Airy sheet?
• RQ3 How does the convergence of the KPZ equation to the KPZ fixed point follow from the convergence of its free energy under scaling?
• RQ4 What role do Baik-Ben Arous-Peche (BBP) laws play in characterizing the KPZ fixed point and enabling convergence proofs?
• RQ5 To what extent can the metric composition and Burke property be used to unify convergence proofs across different stochastic models in 1+1 dimensions?

Key findings

• The O'Connell-Yor polymer model converges in law to the KPZ fixed point under 1-2-3 scaling, with the free energy of the rescaled model converging to the fixed point's distribution.
• The solution to the KPZ equation with deterministic initial condition $ H(0,x) \leq c + b|x| $ converges weakly to the KPZ fixed point, with compact convergence in law for fixed times.
• The KPZ fixed point is uniquely characterized by its distance laws $ \text{Law}(f \cdot \mathcal{S} \cdot g) $, where $ f, g $ are upper semicontinuous functions and $ \mathcal{S} $ is the Airy sheet.
• Convergence of the rescaled models is established via the metric Burke property and tightness of stationary measures, ensuring the limit is well-defined.
• The convergence holds for a broad class of initial conditions, including all classically considered ones, as shown in Corollary 12.
• The method provides a general framework that applies to polymer models and can be extended to other systems, offering an alternative to independent proofs by Quastel and Sharkar (2020).