Determining nodes for extended dissipative systems (Q4382002)

We consider the complex Ginzburg-Landau equation \[ \partial_t u=(1+i \alpha) \Delta u+u-(1+i \beta) u| u|^2, \tag{CGL} \] where \(\alpha\) and \(\beta\) are real parameters. We recall that if \(d=1\) or 2, and if the initial function \(u_0\) is bounded and uniformly continuous, the solution \(u\) of (CGL) is bounded and uniformly continuous for all time. For a positive number \(l\), we denote by \({\mathcal L}_l\) the square lattice of size \(l\). Associated to this lattice, we introduce the set \({\mathcal Q}_l\) of all cubes of size \(l\) with corners on \({\mathcal L}_l\). For \(\rho>0\) and \(Q\) a cube, we will denote by \(\rho Q\) the cube with the same centre as \(Q\) and with size \(\rho\) times the size of \(Q\). Theorem 1. There is a positive number \(l_c= l_c(\alpha, \beta)<1\) such that if \(0<l <l_c\) and \(u_1\) and \(u_2\) are any two solutions of (CGL) as above and such that for all cubes \(Q\in {\mathcal Q}_l\) \[ \lim_{t \to\infty} \Bigl| \int_Q \bigl(u_1(x,t) -u_2(x,t) \bigr)dx \Bigr| =0, \] then \(u_1-u_2\) converges to zero uniformly on every compact subset of \(\mathbb{R}^d\) when \(t\) tends to infinity. Theorem 2. There is a positive number \(l_c= l_c(\alpha, \beta)<1\) such that if \(0<l <l_c\) and \(u_1\) and \(u_2\) are any two solutions of (CGL) as above and such that for all points \(X\in {\mathcal L}_l\) \[ \lim_{t\to \infty} \bigl| u_1(X,t) -u_2(X,t) \bigr| =0, \] then \(u_1-u_2\) converges to zero uniformly on every compact subset of \(\mathbb{R}^d\) when \(t\) tends to infinity. Theorem 3. In dimension 1, there is a positive number \(l_c= l_c(\alpha, \beta)<1\) such that if \(a\) and \(b\) are two points on the real line satisfying \(0<| a-b |<l_c\) and \(u_1\) and \(u_2\) are two solutions of (CGL) such that \[ \lim_{t\to \infty} \bigl(u_1 (a,t)-u_2 (a,t)\bigr) =0 \quad \text{and} \quad\lim_{t\to \infty} \bigl(u_1 (b,t)- u_2(b,t) \bigr)=0, \] then \(u_1-u_2\) converges to zero uniformly on every compact subset of \(\mathbb{R}\) when \(t\) tends to infinity.