Large deviations for a Burgers'-type SPDE (Q1613652): Difference between revisions

A stochastic partial differential equation \[ \frac {\partial u^\varepsilon }{\partial t} = \frac {\partial ^2 u^\varepsilon }{\partial x^2} + f(t,x,u^\varepsilon) + \frac {\partial }{\partial x}g(t,x,u^\varepsilon) + \sqrt { \varepsilon }\sigma (t,x,u^\varepsilon)\frac {\partial ^2 W} {\partial t\partial x}, \tag{1} \] \(t\in [0,T]\), \(x\in [0,1]\), with the Dirichlet boundary conditions \(u(t,0) = u(t,1) = 0\) is considered. In (1), \(W\) denotes the Brownian sheet, \(\varepsilon >0\), \(\sigma \) is a bounded uniformly continuous function, Lipschitz continuous in the third variable, \(g(t,x,v) = g_1(t,x,v) + g_2(t,v)\), and the functions \(f\), \(g_{i}\) are Borel, locally Lipschitz in the third variable, \(f\), \(g_1\) satisfy the linear growth condition, \(g_2\) is of a quadratic growth. (With the choice \(g(t,x,v) = \frac 12 v^2\), the problem (1) reduces to the stochastic Burgers equation.) \textit{I. Gyöngy} [Stochastic Processes Appl. 73, No. 2, 271-299 (1998; Zbl 0942.60058)] proved that, given an initial condition \(u_0\in L^{p}([0,1])\), \(p\geq 2\), there exists a unique mild solution \(u^\varepsilon \) of (1) and \(u^\varepsilon \in C([0,T];L^{p}([0,1]))\) almost surely. It is shown that the laws of \(u^\varepsilon \), \(\varepsilon >0\), obey on \(C([0,T];L^{p}([0,1]))\) a large deviation principle.