No label defined (Q4526430)

The authors consider the vibrating inhomogeneous beam equation of the Euler-Bernoulli type: having the state equation: \(\rho y_{tt}+ (py'')''= 0\), with a mass \(M\) attached to its tip \(x= L\). Here \(p\) stands for \(EI(x)\), where \(E\) is Young's modulus and \(I\) is the moment of inertia of the cross-sectional area about the neutral axis of bending. \(\rho\) is the material density, and \((x,t)\in (0,L)\times (0,\infty)\), with boundary conditions: \(y(0,t)= y'(0, t)= 0\), and also we have the separate tip equations \(Jy'_{tt}(L, t)= m(t)\), and \(My'_{tt}(L, t)- (py'')'(L, t)= f(t)\), \(f\) is the force, and \(m\) the moment applied at the tip. \(J\) is the rotary inertia of the tip mass \(M\). The energy of the beam \(E(t)\) is the sum of potential and kinetic energy, the admissible terms \(f(t)\) and \(m(t)\) are the controls. The system is called exponentially stabilizable if \(\forall t>0\), \(E(t)\leq Ke^{-\omega t}E(0)\) (for an suitable \(\omega\)). For equations with constant coefficients much is known about not only stabilizability but also about optimal control that decreases the total energy at the fastest rate. The authors introduce the well known to the engineers Hilbert space with the inner product given by \[ \langle Y_1,Y_2\rangle= \int^L_0 [py''_1 y''_2+ (\rho v_1v_2)] dx, \] modified by addition of the terms \(J\acute\gamma_1 \acute\eta_2\), or of \(M\acute\gamma_1 \acute\gamma_2\). The authors assume here that either \(M\) or \(J\) are equal to zero, simplifying their analysis. The system given above is reduced to the evolution equation \(dz/dt= Az\), with assumptions of finiteness (\(L^\infty\) norm). A generates a strongly continuous semigroup of contractions, and \(A^{-1}\) is a compact operator. The authors prove the following spectral theorem: Let \(\rho\) and \(p\) be bounded away from zero and be continuously differentiable functions. Then \(\limsup_{\delta\to\infty}\|i\delta- A^{-1}\|< \infty\). The reviewer comments that by using purely engineering arguments, or by taking appropriate limits one can argue that the \(C^1\) requirement is too strict. Bounded variation continuity would suffice, but this would not be an important improvement. Using frequency domain arguments the authors prove that in each case (either \(M= 0\), or \(J= 0\)) the system is exponentially stabilizable. This paper is well written and complements much of the theory developed for the constant coefficients case.