Table of Contents
Fetching ...

Donsker-Varadhan large deviation principle for locally damped and randomly forced NLS equations

Yuxuan Chen, Shengquan Xiang

TL;DR

This work establishes a Donsker–Varadhan large deviation principle for randomly forced PDEs with damping and degenerate noise acting on finite modes. It introduces a general LDP criterion for random dynamical systems, built on irreducibility and a fixed-rate squeezing condition, and overcomes the challenge of a non-vanishing squeezing rate with a novel bootstrap argument for Lipschitz bounds on Feynman–Kac semigroups. The authors verify the criterion in three settings: a nonlinear Schrödinger equation on $ ext{1D torus}$, a nonlinear wave equation, and the two-dimensional Navier–Stokes system with degenerate noise, obtaining uniform LDPs for empirical measures on the support of the unique invariant measure $ mu_*$. The results pave the way for LDP analyses of broader dissipative PDEs with degenerate forcing and offer a pathway to studying observables via the contraction principle. They also discuss the limitations related to initial data outside $ ext{supp}( mu_*)$ and outline future work on hyperbolic equations with white-in-time noise.

Abstract

We study large deviations from the invariant measure for nonlinear Schrödinger equations with colored noises on determining modes. The proof is based on a new abstract criterion, inspired by [V. Jakšić et al., Comm. Pure Appl. Math., 68 (2015), 2108-2143]. To address the difficulty caused by fixed squeezing rate, we introduce a bootstrap argument to derive Lipschitz estimates for Feynman-Kac semigroups. This criterion is also applicable to wave equations and Navier-Stokes system.

Donsker-Varadhan large deviation principle for locally damped and randomly forced NLS equations

TL;DR

This work establishes a Donsker–Varadhan large deviation principle for randomly forced PDEs with damping and degenerate noise acting on finite modes. It introduces a general LDP criterion for random dynamical systems, built on irreducibility and a fixed-rate squeezing condition, and overcomes the challenge of a non-vanishing squeezing rate with a novel bootstrap argument for Lipschitz bounds on Feynman–Kac semigroups. The authors verify the criterion in three settings: a nonlinear Schrödinger equation on , a nonlinear wave equation, and the two-dimensional Navier–Stokes system with degenerate noise, obtaining uniform LDPs for empirical measures on the support of the unique invariant measure . The results pave the way for LDP analyses of broader dissipative PDEs with degenerate forcing and offer a pathway to studying observables via the contraction principle. They also discuss the limitations related to initial data outside and outline future work on hyperbolic equations with white-in-time noise.

Abstract

We study large deviations from the invariant measure for nonlinear Schrödinger equations with colored noises on determining modes. The proof is based on a new abstract criterion, inspired by [V. Jakšić et al., Comm. Pure Appl. Math., 68 (2015), 2108-2143]. To address the difficulty caused by fixed squeezing rate, we introduce a bootstrap argument to derive Lipschitz estimates for Feynman-Kac semigroups. This criterion is also applicable to wave equations and Navier-Stokes system.

Paper Structure

This paper contains 16 sections, 14 theorems, 89 equations.

Key Result

Proposition 2.1

Under the above settings, suppose that hypotheses $(\mathbf{I})$ and $(\mathbf{C})$ are satisfied. Then the Markov process $(x_n,\mathbb{P}_x)$ defined by RDS admits a unique invariant measure $\mu_*\in\mathcal{P}(X)$. Moreover, $\operatorname{supp}(\mu_*)=X$, and there exist constants $C,\gamma>0$

Theorems & Definitions (35)

  • Example 1.1: Parabolic equation with degenerate noise
  • Example 1.2: Hyperbolic equation
  • Proposition 2.1
  • Theorem 2.2
  • Remark 2.3
  • Definition 2.4: Attainable set
  • proof : Proof of the Main Theorem
  • proof : Proof of the Main Theorem
  • Definition 2.5: $\Gamma$-type domain
  • Theorem 2.6
  • ...and 25 more