Two-endpoint boundary control of vibrations described by a finite-energy generalized solution of the wave equation (Q5951238)

The author continues his study of a vibrating string (wave) equation \(u_{tt}- u_{xx}= 0\), \(0\leq t\leq T\). Displacements and velocities of the string are equal to \(\varphi(x)\) and \(\psi(x)\), respectively at the time \(t=0\) and \(\varphi_1(x)\) and \(\psi_1(x)\) at \(t= T\). The problem posed here is to describe boundary controls. The author states: ``In the class of finite energy generalized solutions of the wave equation given any \(T>0\), we intend to obtain necessary and sufficient conditions for the existence of boundary controls \(u(0,t)= \mu(t)\) and \(u(\ell, t)=\nu(t)\) bringing vibration to the state \(\{u(x,T)= \varphi_1(x)\), \(u(x,T)= \psi_1(x)\}\) at time \(t=T\), and to find explicit closed form expressions for these controls.'' It is this result stated in the last part of this sentence that is missing in many articles concerning the control of the wave equation, beginning with D. L. Russel's MRC report in 1965, and the reviewer's publication in 1968, where these details were missing for even the two-dimensional case. In a correspondence between the reviewer and William Prager some time in the seventies, the reviewer expressed an opinion that the time interval should be expanded to \(2\ell\) and the controls for the case discussed here should be of d'Alambert's form \(f(x+ t)\) and \(g(x-t)\), to which Professor Prager offered a simple counterexample, implying that since the length \(\ell\) appears in Green's function for this problem, then \(\ell\) should appear in the control as well. The author finds the exact detail that was missing. \(u(x,t)\) is equal to the sum of a control function of \(x-t\) plus a function of \((x+ t)-\ell\). If the time interval is extended to \(0< T< 2\ell\), then the formula is slightly more complex, but Prager's implication holds. The author also proves necessary conditions that must be obeyed by the functions defining the boundary terms for the existence of a \(W^1_2\) unique solution of this boundary control problem. The author has a long record of research in applied mathematics and mechanics beginning in the 1950's. The present article is an example of his ability to solve long-standing problems using classical analysis combined with attention to details.