Adaptive stabilization of infinite-dimensional semilinear second-order systems (Q4408790)

The author of this paper investigates adaptive stabilization of infinite-dimensional undamped semilinear second-order systems with unbounded input and output operators of the form \[ {d^2z\over dt^2}+ Az+ f(z)= Bu(t)\quad\text{on }(0,T),\tag{1} \] \[ z(0)= z_0,\quad {dz(0)\over dt}= z_1,\tag{2} \] \[ y(t)= B^*\dot z(t),\tag{3} \] where \(u\in L^2(0,T)\), \(z_0\in V\), \(z_1\in H\), \(V\) and \(H\) are Hilbert spaces so that \(V\subset H\equiv H'\subset V'\), where the injections are continuous and each space is dense in the following one. An adaptive stabilizer for the system (1)--(3) is constructed by the concept of high-gain output feedback. An adaptive controller \[ \begin{aligned} u(t)&= -k(t) y(t),\\ \dot k(t) &= ry^2(t),\;k(0)= \kappa\geq 0,\;r>0\end{aligned} \] is applied to the system (1)--(3) and the resulting closed-loop system becomes \[ {d^2z\over dt^2}+ k(t) BB^*{dz\over dt}+ Az+ f(z)= 0\quad\text{on }(0,T),\tag{4} \] \[ z(0)= z_0,\qquad {dz(0)\over dt}= z_1, \] \[ \begin{cases} \dot k(t)= ry^2(t),\quad k(0)\geq 0,\;r>0,\\ y(t)= B^*\dot z(t).\end{cases}\tag{5} \] In order to prove the asymptotic stability of the closed-loop system (4)--(5) the author defines the following energy-like functions: \[ \begin{aligned} E(t) &= \textstyle{{1\over 2}} (\dot z(t),\dot z(t))_H+ \textstyle{{1\over 2}} (A^{1/2} z(t), A^{1/2}z(t))_H+ F(z(t)),\\ V(t) &= E(t)+ \varepsilon g(t),\quad \varepsilon> 0.\end{aligned} \] The existence of the energy-like function \(E(t)\) with \(\dot E(t)= -k(t)y^2(t)\) and the multiplier \(g(t)\) satisfying the assumptions \[ | g(t)|\leq C_1 E(t)\quad\text{for all }t\geq 0, \] \[ \dot g(t)\leq -C_2E(t)+ \phi(k(t)) y^2(t)\quad\text{for all }t\geq 0, \] where \(C_1> 0\), \(C_2>0\) and the function \(\phi\) is such that \(|\phi(k)|< \infty\) if \(| k|< \infty\), is essential to adaptively stabilize the second-order systems.