Invariant measures for the open KPZ equation: an analytic perspective

Alexander Dunlap, Yu Gu, Tommaso RosatiarXiv 2025 · DOI · arXiv

TL;DR

This paper provides the first direct stochastic analytic proof for the invariant measures of the open Kardar-Parisi-Zhang (KPZ) equation on a finite interval. By moving away from discrete model approximations and leveraging the theory of regularity structures and time-reversal symmetry, the authors demonstrate how boundary conditions transform simple Gaussian noise into complex, non-Gaussian steady states.

What Problem Does This Paper Address?

The KPZ equation is a fundamental stochastic partial differential equation (SPDE) used to describe the non-equilibrium growth of interfaces, such as the edge of a spreading fire or a growing crystal. When this equation is "open"—meaning it is defined on a finite interval with specific slopes (Neumann boundary conditions) at the ends—determining its "steady state" or invariant measure becomes notoriously difficult.

For years, the only way to describe these invariant measures was through "integrable probability." Researchers would find a discrete model (like a particle system on a lattice) that approximated the KPZ equation, solve for its steady state, and then take a limit to reach the continuum. While successful, this approach is indirect and often relies on specific parameter regimes. The authors of this paper tackle the challenge of proving these measures directly from the SPDE itself, providing a robust, analytic proof that works for all boundary parameters.

Methods

The authors employ a sophisticated toolkit from modern stochastic analysis:

  • Singular Change of Measure: They treat the inhomogeneous boundary conditions (the specific slopes at the ends) as a singular perturbation of the noise. Using the Cameron–Martin theorem, they tilt the underlying probability measure to account for the boundary effects.
  • Time Reversal: They exploit a "skew-symmetry" property. In the simplest case (zero slopes), the time-reversed version of the process looks like the forward process but with the nonlinearity flipped. This symmetry is a crucial ingredient in uncovering the martingale structure of the solution.
  • Regularity Structures: Because the KPZ equation is "singular" (its solutions are extremely rough, like Brownian motion), standard calculus doesn't apply. The authors use Martin Hairer’s theory of regularity structures to perform a "local expansion" of the solution near the boundaries.
  • Boundary Layer Analysis: The proof reduces the problem to analyzing the KPZ nonlinearity in a very thin layer near the edges. They establish a Central Limit Theorem (CLT) for the time-integrated nonlinearity in this layer to complete the proof of invariance.

Datasets

Not discussed in the paper. As a theoretical mathematics paper, it focuses on rigorous proofs and symbolic derivations rather than empirical data or benchmarks.

Key Insights and Ideas

  • Boundary Conditions as Forcing: A key heuristic insight is that inhomogeneous Neumann boundary conditions can be mathematically interpreted as Dirac delta forcing (point-source hits) at the boundaries.
  • The Martingale Connection: The authors uncover a hidden martingale structure. They show that a specific functional of the solution—related to the predicted Radon–Nikodym derivative of the invariant measure—is a martingale under the backward filtration.
  • Robustness Across Regimes: Unlike discrete models that might only be solvable for certain parameters, this analytic method is "robust." It treats the boundary parameters $u$ and $v$ as general variables, meaning the proof holds regardless of the specific physical regime the system is in.
  • Non-Gaussianity is Local: The "aha moment" is that the non-Gaussian nature of the invariant measure is driven by the physics happening in the "boundary layer." Once you correctly account for the integrated boundary flux, the rest of the proof follows from the known Gaussian behavior of the bulk.

Key Results

  • Theorem 1.1: The authors rigorously prove that the invariant measure for the open stochastic Burgers equation is absolutely continuous with respect to spatial white noise, with the Radon–Nikodym derivative exactly matching the formula conjectured by previous integrable probability works.
  • CLT for Boundary Flux: They successfully establish a Central Limit Theorem for the time-integrated nonlinearity near the boundary (Proposition 3.3), showing it converges to a Gaussian random variable independent of the "bulk" noise.
  • Unified Proof: The proof proceeds in a uniform manner for all boundary parameters $(u, v)$, avoiding the need for analytic continuation which was a hurdle in previous methodologies.

Limitations and Future Work

  • Two-Layer Structure: While the proof is complete, the authors note they do not explicitly utilize the "two-layer" Gibbs measure structure found in some discrete models. Exploring how their analytic method relates to this structure could be a fruitful path.
  • Process-Level Convergence: The authors suggest that with additional effort, one could extend their results from a CLT for the integrated flux to a full process-level convergence to a drifted Brownian motion.
  • Higher Dimensions: The KPZ equation in dimensions higher than one remains one of the greatest challenges in the field. While this paper provides a definitive look at the 1D open case, the higher-dimensional boundary problem is still largely unexplored.

Summary updated

Abstract

The ergodic theory of the open KPZ equation has seen significant progress in recent years, with explicit invariant measures described in a series of works by Corwin--Knizel, Barraquand--Le Doussal, and Bryc--Kuznetsov--Wang--Wesołowski. In this paper, we provide a stochastic analytic proof of the formula for the invariant measures. Our approach starts from the Gaussian invariant measure for the case of homogeneous boundary conditions. We approximate the inhomogeneous problem by a homogeneous one with a singular boundary potential. Using tools including change of measure, time reversal for Markov processes, and Itô's formula, we then reduce the problem to analyzing the KPZ nonlinearity in a thin boundary layer. Finally, using the theory of regularity structures, we establish a central limit theorem for the time-integrated nonlinearity near the boundary, which completes the proof of the invariance. Although it is known that different boundary parameters give rise to distinct physical regimes for the invariant measures, our method is robust and does not rely on any particular choice of boundary parameters.
BibTeX
@misc{dunlap2025invariant,
  author = {Alexander Dunlap and Yu Gu and Tommaso Rosati},
  title = {{Invariant measures for the open KPZ equation: an analytic perspective}},
  year = {2025},
  eprint = {2512.03328},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2512.03328},
  howpublished = {arXiv},
}
Figure