On the stochastic Korteweg-de Vries equation driven by white noise (Q5933790)

Consider, the stochastic Korteweg-de Vries equation of the form \[ U_t-6UU_x+U_{xxx}=\xi(t),\tag{1} \] where \(\xi(t)=\xi(t;\omega)\) is a random process, with \(\omega\in\Omega\), \((\Omega,{\mathcal B},P)\) a probability space. Basic to our approach is a relation discovered by Miura between solution of (1) and the usual Kortetweg-de Vries equation \[ u_t-6uu_x+u_{xxx}=0.\tag{2} \] Miura observed that if \(u(x,t)\) satisfies (2), and we define \[ U(x,t,\omega)=u\left(x+6\int^t_0 W(s,\omega)ds,t\right)+W(t,\omega), \] where \(W(t,\omega)=\int^t_0\xi(s,\omega)ds,\) then \(U(x,t,\omega)\) satisfies (1). Conversely, if \(U(x,t,\omega)\) satisfies (1), and we define \[ u(x,t,\omega)=U\left(x-6 \int^t_0 W(s,\omega)ds,t,\omega\right)-W(t,\omega), \] then, for every \(\omega\in\Omega\), \(u(x,t,\omega)\) satisfies (2). The author proves the following main result: Let \(U(x,t,\omega)\) satisfy the stochastic KdV equation (1). Suppose (i) \(U_x,U_{xx},\dots\to 0\), and \(U-W(t)\to 0\) as \(|x|\to 0\), for each \(t\geq 0\) and a.e. \(\omega\), (ii) \({\mathcal F}(u)\) is a conserved density for the KdV equation (2), provided \(u,u_x,u_{xx},\dots\to 0\) as \(|x|\to\infty\), and, (iii) \(\int^\infty_{-\infty}E({\mathcal F}(U(x,0)))dx\) exists and is finite. Then \(\int^\infty_{-\infty} E({\mathcal F}(U(x,t)-W(t))dx\) is independent of \(t\).