Exact null controllability of structurally damped and thermoelastic parabolic models (Q1287018)

The authors consider null controllability for a non-classical parabolic PDE with distributed control. The problem is posed in a Hilbert space \(X\). Two classes of parabolic-like equations are: 1) Structurally damped second-order (abstract) equations: \(w_{tt}+ Sw+ \rho S^\alpha w_t= u\), where \(\rho>0\), \(S\) is a strictly positive spatial selfadjoint and unbounded partial differential operator, which in the simplest case could be \(-d^2/dx^2\), \(S^\alpha\) is a corresponding fractional order operator, with \({1\over 2}\leq\alpha<1\), \(u\in L^2\{[0, T]\}\) is the control. The initial conditions are: \(w(0)= w_0\), \(w_t(0)= w_1\). This equation can be rewritten in variable \(z= \{w,w_t\}\): using the matrix operator: \[ A_{\rho\alpha}= \left[\begin{matrix} 0 & I\\ -S & -\rho S^\alpha\end{matrix}\right] \] as a first-order equation: \(z_t= A_{\rho\alpha}z+ Bu\). 2) An abstract thermoelastic equation with hinge boundary conditions and Dirichlet thermal conditions: \(y_t= {\mathcal A}y+ Bu\), where \[ {\mathcal A}= \left[\begin{matrix} 0 & A & 0\\ -A & 0 & A\\ 0 & -A & -A\end{matrix}\right], \] and the operator \(B= [0\;I\;0]\), or \([0\;0\;I]\), depending on whether mechanical or thermal control is applied. These mathematical systems are called abstract, since they do not model some specific physical system, but may fit several approximations to controlled mechanical systems. The authors remark that 1) was investigated by \textit{S. A. Avdonin} and \textit{S. A. Ivanov} in their monograph [Families of exponentials. The method of moments in controllability problems for distributed parameter systems. Cambridge University Press (1995; Zbl 0866.93001)], with \(S= -d^2/dx^2\). They showed that for \({1\over 2}\leq\alpha\leq 1\) the system is not exactly null-controllable when \(u\in L^2\{[0, T]\}\) is a scalar. However, assuming that \(u\in L^2\{[0, T],X\}\) and \(\alpha< 1\), for distributed controls, the authors succeed in proving null-controllability. The authors use strongly continuous semigroup theory, making use of duality, factorization of the operators, followed by diagonalization, and some new and rather subtle inequalities. They maintain the very high standard of mathematical sophistication developed in their past studies of control theory for mechanical systems.