The 2D nonlinear stochastic heat equation: pointwise statistics and the decoupling function
Abstract
We consider a two-dimensional stochastic heat equation with Gaussian noise of correlation length $\rho^{1/2} \ll 1$ and strength $\sigma(v)/\sqrt{\log\rho^{-1}}$ depending nonlinearly on the solution $v$. The first author and Gu showed that, if $\operatorname{Lip}(\sigma)$ lies below a critical value, the one-point statistics of $v$ converge in distribution as $\rho\to 0$ to the law of a random variable expressed in terms of a forward-backward stochastic differential equation (FBSDE). In this paper, we show (for a slightly different smoothing mechanism) that this result can hold even when $\operatorname{Lip}(\sigma)$ exceeds the critical value. More precisely, the FBSDE characterizes the limiting one-point statistics of the SPDE so long as the square root of the decoupling function of the FBSDE remains Lipschitz. In addition, we extend the results to coupled systems of SPDEs and provide a quantitative rate of convergence.
- Publication:
-
arXiv e-prints
- Pub Date:
- August 2023
- DOI:
- arXiv:
- arXiv:2308.11850
- Bibcode:
- 2023arXiv230811850D
- Keywords:
-
- Mathematics - Probability;
- Mathematics - Analysis of PDEs;
- 60H15 (Primary);
- 35R60;
- 60H10 (Secondary)
- E-Print:
- 82 pages