An application of inertial manifold theory to boundary stabilization of semilinear diffusion systems (Q1910043)

The authors consider a diffusion system described by the equations: \[ \begin{aligned} \partial z (t,x)/ \partial t &= (\Delta + \gamma) z(t,x) + f(z (t,x)), \qquad (t,x) \in (0, \infty) \times \Omega, \quad \gamma > 0,\\ \partial z (t, \xi)/ \partial \nu &= g(\xi) u(t), \qquad (t, \xi) \in (0, \infty) \times \Gamma , \quad z(0,x) = z_0 (x),\;x \in \Omega.\end{aligned} \tag{1} \] The output of the system is given by: \(Y(t) = \int_\Omega w(x) z(t,x) dx\). Here \(z\) belongs to \(H^2 (\Omega)\). An abstract version of the system (1) is: \(Az = (\Delta + \gamma) z + f(z)\), \(z \in H^2 (\Omega)\), with \(\partial z/ \partial \nu = 0\) on \(\Gamma\). Here \(A\) is self-adjoint with a compact resolvent. The eigenvalues of \(A\) are given in pairs \(\lambda_i = i^2 - \gamma\), and the eigenfunctions are \(\varphi_0 = 1/ \pi^{1/2}\), \(\varphi_i = (2 \pi)^{1/2} \cos ix\), \(i = 1,2, \dots\) An operator \(A_c = A + c\), \(c > 0\), is introduced and fractional powers of \(A_c\), specifically \(A_c^{1/4 + \varepsilon}\), \(A_c^{- 1/4 - \varepsilon}\) are defined in appropriate Hilbert spaces. The reviewer comments that the use of fractional order operators of the type \(A^{1/4 + \varepsilon}\), where \(A\) is a second order linear differential operator, has been pioneered by Gingold, Balakrishnan and others in the classical theory of ordinary differential equations, but the original proofs of the containments of their domains go back to the work of T. Kato in the 1950-s. In the present paper, the domain \(\Omega\) is one dimensional. Absence of the spectral gap condition seems to prevent derivation of much more general results. Having decomposed the original operator, the authors proceed to define groups and semigroups associated with various perturbations of the system. Their aim is to establish the existence of an invariant, globally attractive, finite dimensional Lipschitz manifold, (that is an inertial manifold) such that closed loop stabilization of the system reduces to the use of a Sakawa-type controller on the flow on this manifold. This paper deserves a careful study. There are numerous fine details and theoretical arguments which are impossible to include in a brief review. For example, the use of a Sakawa-type controller allows the authors to prove existence of stabilization, and to estimate the appropriate Lipschitz constant. Then they are free to use the inertial manifold theory as discussed by \textit{S. N. Chow} and \textit{K. Lu} (J. Differential Equations 74, No. 2, 309-353 (1988; Zbl 0691.58034).