\(C^ 1\) approximations of inertial manifolds for dissipative nonlinear equations (Q1913661)

From MaRDI portal
scientific article
Language Label Description Also known as
English
\(C^ 1\) approximations of inertial manifolds for dissipative nonlinear equations
scientific article

    Statements

    \(C^ 1\) approximations of inertial manifolds for dissipative nonlinear equations (English)
    0 references
    0 references
    0 references
    9 July 1996
    0 references
    A class of nonlinear dissipative partial differential equations which have inertial manifolds is studied. The equations are of the form: \[ {du\over dt}+ Au+ R(u)= f,\quad u(0)= u_0.\tag{1} \] Here \(A\) is an unbounded positive selfadjoint linear operator with compact inverse defined on \(D(A)\subset H\) and \(H\) is Hilbert space. Thus there exists an orthonormal basis \(\{\varphi_j\}\) of \(H\) consisting of eigenvectors of \(A\), \(A\varphi_j= \lambda_j \varphi_j\), and \(\lambda_j\) satisfy \(0< \lambda_1\leq \lambda_2\leq\cdots\leq \lambda_j\to \infty\) as \(j\to \infty\). Moreover, \(R(u)\) is assumed to be a differentiable map from \(D(A)\) into \(D(A^{1- \beta})\) for some \(\beta\), \(0\leq \beta\leq {1\over 2}\) and satisfies some growing conditions. The function \(f\) is supposed to be in \(H\). It is supposed that for every \(u_0\in D(A)\), equation (1) has a unique global solution which is denoted \(S(t) u_0\) and for all \(t\geq 0\), \(S(t) u_0\in D(A)\) and is continuous in both variables. In this paper, the linear operator \(A\) is assumed to have large gaps in its spectrum. The reduction of the partial differential equation to the inertial manifold yields the inertial form. More precisely, denoting by \(P\) the orthogonal projection on the space \(H\) onto \(PH= \text{span}\{\varphi_1,\dots, \varphi_M\}\), \(Q= I- P\), setting \(p= Pu\), \(q= Qu\), the evolution equation (1) is equivalent to the system \[ {dp\over dt}+ Ap+ PR(p+ q)= Pf,\quad {dq\over dt}+ Aq+ QR(p+ q)= Qf.\tag{2} \] If the inertial manifold is given as a graph of a Lipschitz function \(\Phi: PH\to QD(A)\) then on this manifold the solutions of (1) are of the form \(u(t)= p(t)+ \Phi(p(t))\). In this case the inertial form is given by \(dp/dt+ Ap+ PR(p+ \Phi(p))= Pf\), \(p\in PH\). In this paper, it is shown that a Galerkin approximation of equation (1) is a small \(C^1\) perturbation of (2), the inertial form, provided the system is taken sufficiently large.
    0 references
    dissipative nonlinear equations
    0 references
    inertial manifolds
    0 references
    inertial form
    0 references
    Galerkin approximation
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references