Nikolay Barashkov,Francesco C. De Vecchi,Immanuel Zachhuber
Abstract:We study the Gaussian measure whose covariance is related to the Anderson Hamiltonian operator, proving that it admits a regular coupling to the (standard) Gaussian free field exploiting the stochastic optimal control formulation of Gibbs measures. Using this coupling we define the renormalised powers of the Anderson free field and we prove that the associated quartic Gibbs measure is invariant under the flow of a nonlinear wave equation with renormalised cubic nonlinearity.
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is the invariant measure problem regarding the nonlinear Anderson wave equation. Specifically, the authors studied a Gaussian measure whose covariance is related to the Anderson Hamiltonian operator, with the goal of proving that this measure is invariant under the flow of the nonlinear wave equation. This nonlinear wave equation has the form:
\[
\partial_t^2 u(t,x) + H_\omega u(t,x) = - u^3(t,x), \quad t \in \mathbb{R}, \quad x \in \mathbb{T}^2, \quad \omega \in \Omega
\]
where \(H_\omega\) is the Anderson Hamiltonian formally defined as a singular Schrödinger operator:
\[
H_\omega u(t,x) = -\Delta_{\mathbb{T}^2} u(t,x) + \xi(\omega, x) u(x,t)
\]
Here, \(\xi(\cdot, x) : \Omega \to S'(\mathbb{T}^2)\) is a Gaussian white noise defined on a certain probability space.
The authors' goal is to prove that the formal measure:
\[
\exp\left(-\frac{1}{2}(\phi, H_\omega \phi) dx - \frac{1}{2} \int |\partial_t \phi|^2 dx - \frac{1}{4} \int \phi^4 dx\right) D(\phi, \partial_t \phi)
\]
is invariant under the flow of this nonlinear wave equation, where \(D(\phi, \partial_t \phi)\) is the Lebesgue measure on some (nonexistent) function space.
To achieve this goal, the authors need to deal with several key technical problems, including:
1. **Regular coupling**: Construct a regular coupling between the standard Gaussian free field and the Anderson Gaussian free field through the formulation method of stochastic optimal control.
2. **Renormalization**: Renormalize the powers of the Anderson free field and prove that the renormalized quartic Gibbs measure is invariant under the flow of the nonlinear wave equation.
3. **Low - regularity handling**: Due to the low regularity of the measure support, it is necessary to Wick - order the cubic nonlinear term, that is, subtract a divergent linear counter - term to make it well - defined.
The solutions to these problems involve not only the study of Gaussian measures and Wick powers, but also the strict construction of the Anderson Hamiltonian and the analysis of its properties. Through these technical means, the authors finally prove the invariance of the target measure under the nonlinear wave equation.