Space-time stationary solutions for the Burgers equation

TL;DR

This paper constructs unique space-time stationary solutions for the 1D Burgers equation with random Poissonian forcing, establishing their role as attractors and stationary distributions for the associated Markov process.

Contribution

It introduces a novel construction of global solutions without periodicity assumptions, using action-minimizing curves, and characterizes their convergence and uniqueness.

Findings

Existence of unique global solutions for any prescribed average velocity.

These solutions act as one-point random attractors.

Convergence of the Markov process distribution to the stationary distribution.

Abstract

We construct space-time stationary solutions of the 1D Burgers equation with random forcing in the absence of periodicity or any other compactness assumptions. More precisely, for the forcing given by a homogeneous Poissonian point field in space-time we prove that there is a unique global solution with any prescribed average velocity. These global solutions serve as one-point random attractors for the infinite-dimensional dynamical system associated to solutions to the Cauchy problem. The probability distribution of the global solutions defines a stationary distribution for the corresponding Markov process. We describe a broad class of initial Cauchy data for which the distribution of the Markov process converges to the above stationary distribution. Our construction of the global solutions is based on a study of the field of action-minimizing curves. We prove that for an arbitrary value of the average velocity, with probability 1 there exists a unique field of action-minimizing curves initiated at all of the Poissonian points. Moreover action-minimizing curves corresponding to different starting points merge with each other in finite time.