Space-time invariant measures, entropy, and dimension for stochastic Ginzburg-Landau equations (Q1348983)

Equations of the form \[ du=((1+i\alpha) \Delta u+u- (1+i\beta) |u|^{2q}u) dt+\xi(x)dw(t),\;u(x,t)\in\mathbb{C}, x\in\mathbb{R}^d, t\geq 0, \alpha,\beta \in\mathbb{R}, \tag{1} \] where \(\Delta\) is the \(d\)-dimensional Laplacian, \(w(t)\) is a Wiener process, are considered. Uniform local Sobolev spaces \(H^m_{\text{ul}}\) and the random attractor \(A_\omega\) are defined. For equation (1) the spatial densities of topological entropy \((h_{\text{top}})\), of measure-theoretic entropy \((h_\mu)\), and of upper box-counting dimension of \(A_\omega (d_{\text{up}})\), are defined. \[ \text{Hypothesis: }d\leq 2,q\geq{1\over 2},-(1+ \alpha\beta) <|\alpha- \beta|{\sqrt{2q+1} \over q},|\beta |\leq {\sqrt{2q+1} \over q}\tag{2} \] The main results of the paper: (1) If (2) holds, then (1) with initial data \(u_0\in C^\infty_b\) has a unique solution \(u(x,t)= u_t(x)\). \(\forall p\geq 1\), \(m\in\mathbb{N}\), \(\exists B_{p,m} <\infty\) such that \(\forall t>t_0 (u_0)\) \(E\|u_t\|^p_{H^m_{\text{ul}}}\leq B_{p,m}\). (2) There exists at least one invariant measure \(\mu\) for (1). This measure is homogeneous in \(x\) and its support is contained in \(\bigcap_{m\geq 0}H^m_{\text{ul}}\). (3) \(\exists \gamma <\infty: h_\mu\leq h_{\text{top}} \leq\gamma d_{\text{up}} <\infty\).