Strong invariance and noise-comparison principles for some parabolic stochastic PDEs (Q516123)

The authors consider the partial differential equation \[ \frac{\partial}{\partial t} u_t (x) = ( {\mathcal{L}} u_t ) (x) + \sigma (u_t(x)) \xi(t,x), \] \(t>0, x \in \mathbb{R}\) and \(u_0(x)=1\) for all \(x \in \mathbb{R}\), where the forcing term is a space-time white noise. The operator \( {\mathcal{L}}\) is the fractional Laplacian on \(\mathbb R\) of order \(\alpha\). \(u_t(x)\) can be seen as the conditional expected density of a branching particle system at space-time point \((t,x)\) given the white noise \(\xi\) and where the particles move in \(\mathbb{R}\) as an \(\alpha\)-stable Lévy process. On the other hand, they consider the interacting particle system \[ d U_t (x) = ( {\mathcal{S}} U_t ) (x) dt + \sigma (U_t(x)) dB_t(x), \] \(t>0, x \in \mathbb{Z}\) and \(U_0(x)=1\) for all \(x \in \mathbb{Z}\). Here, \({\mathcal{S}}\) denotes the generator of the rate 1 continuous-time random walk on \( \mathbb{Z}\) and \(dB_t(x)/dt \) is a space-time white noise on \(\mathbb{R}_+ \times \mathbb{Z}\). It is shown that if one rescales the space in the above particle system and then speeds up the rate of the jumps of the underlying walks, the system converges (in a strong sense) to the solution of the initial stochastic partial differential equation. So, the authors establish an infinite dimensional invariance principle. As a consequence, they get comparison inequalities for product moments of the stochastic heat equation with different nonlinearities.