Effective dynamics of interfaces for nonlinear SPDEs driven by multiplicative white noise

TL;DR

This work develops a rigorous reduction framework for nonlinear SPDEs with multiplicative white noise by projecting dynamics onto an approximate slow manifold, yielding a finite-dimensional SDE for the interface parameter $h$ coupled to an infinite-dimensional orthogonal remainder. It proves that the full SPDE is equivalent to the coupled system $(h,v)$ under an invertibility condition on the matrix $A(h,v)$, enabling analysis of metastability and interface motion. The framework is demonstrated on four canonical models—the stochastic damped wave, Allen–Cahn, nonlinear Schrödinger, and Swift–Hohenberg equations—producing reduced dynamics for $h$ that exhibit noise-induced drift, diffusion along the manifold, and long-lived metastable states, with numerical simulations validating the theory. By clarifying how multiplicative noise shapes interface evolution, the results provide a practical toolkit for predicting rare events and pattern evolution in driven systems, and they point to future work on colored/non-Gaussian noise, higher-dimensional interfaces, rigorous convergence proofs, and data-driven coarse-graining.

Abstract

In the present work, we investigate the dynamics of the infinite-dimensional stochastic partial differential equation (SPDE) with multiplicative white noise. We derive the effective equation on the approximate slow manifold in detail by utilizing a finite-dimensional stochastic differential equation (SDE) describing the motion of interfaces. In particular, we verify the equivalence between the full SPDE and the coupled system under small stochastic perturbations. Moreover, we apply our results to effective dynamics of stochastic models with multiplicative white noise, illustrated with four examples on the stochastic damped wave equation, the stochastic Allen-Cahn equation, the stochastic nonlinear Schrödinger equation and the stochastic Swift-Hohenberg equation.

Figures (4)

• Figure 1: (a) A time series comparison (deterministic vs. stochastic), and a histogram of the final state $h(T)$; (b) A phase diagram showing the deterministic drift and sample paths. Damping coefficient $\gamma=10$; Linear damping rate $a=\pi^2/\gamma$; Nonlinear damping coefficient $b=3/\gamma$; Noise intensity $\varepsilon=0.5$; Initial condition $h_0 = 0.1$; Total simulation time $T=50$; Time step $dt=0.01$.
• Figure 2: Front position $h(t)$ in stochastic Allen-Cahn equation. Domain length $L=20$; Noise intensity $\varepsilon=0.1$; Initial front position $h_0=L/2$; Total time $T=10^4$; Time step $dt=0.1$.
• Figure 3: Simulated trajectory of $h(t)$ from the reduced SDE, showing noise-driven diffusion with drift. The soliton maintains positional coherence over long times despite stochastic perturbations. Noise intensity $\varepsilon=0.1$; Initial position $h_0=0$; Total time $T=10^3$; Time step $dt=0.1$.
• Figure 4: Trajectory of $h(t)$ from \ref{['ISH']}, the amplitude fluctuates near $h_{\ast}\approx0.183$ (dashed line), with rare escapes consistent with exponential metastability; Bifurcation parameter $\delta=0.1$; Noise intensity $\varepsilon=0.05$; Initial condition (stable equilibrium) $h_0=\sqrt{\delta/3}$; Total time $T=10^3$; Time step $dt=0.1$.

Theorems & Definitions (12)

• Definition 2.1
• Remark 2.1
• Definition 2.2
• Definition 2.3
• Lemma 2.1
• proof
• Theorem 2.1
• proof
• Remark 2.2
• Corollary 2.1