Stochastic Variational Inequalities and Applications to the Total Variation Flow Perturbed by Linear Multiplicative Noise

Abstract

In this work, we introduce a new method to prove the existence and uniqueness of a variational solution to the stochastic nonlinear diffusion equation $dX(t)={\rm div} [\frac{\nabla X(t)}{|\nabla X(t)|}]dt+X(t)dW(t) in (0,\infty)\times\mathcal{O},$ where $\mathcal{O}$ is a bounded and open domain in $\mathbb{R}^N$, $N\ge 1$, and $W(t)$ is a Wiener process of the form $W(t)=\sum^\infty_{k=1}\mu_k e_k\beta_k(t)$, $e_k \in C^2(\bar\mathcal{O})\cap H^1_0(\mathcal{O}),$ and $\beta_k$, $k\in\mathbb{N}$, are independent Brownian motions. This is a stochastic diffusion equation with a highly singular diffusivity term and one main result established here is that, for all initial conditions in $L^2(\mathcal{O})$, it is well posed in a class of continuous solutions to the corresponding stochastic variational inequality. Thus one obtains a stochastic version of the (minimal) total variation flow. The new approach developed here also allows to prove the finite time extinction of solutions in dimensions $1\le N\le 3$, which is another main result of this work. Keywords: stochastic diffusion equation, Brownian motion, bounded variation, convex functions, bounded variation flow.