The wave equation with a boundary control in the case of elastic fixing. I (Q1874905)

The paper under review is part of a sequence of publications by the author and by Professor Il'in on the study of the classic wave equation \(u_{tt}- u_{xx}= 0\), with \(0< t\leq T\), and \(0\leq x\leq\ell\). Most analysts probably would shrug their shoulders saying: there is the d'Alambert solution \[ \vartheta(x)= {1\over 2} \Biggl[(\varphi(x+ t)+ \varphi(x- t))+ \int^{x+ t}_{x-t} \psi(x)\Biggr], \] where \(\varphi(x)= u(x,0)\) and \(\psi(x)= u_t(x,0)\), which belong, respectively to Sobolev spaces \(W^2_2\) and \(W^1_2\). What else can be said about the one-dimensional case? It turns out that a lot can be said that is not known when discussing the possible (admissible) boundary controls of this one-dimensional wave. In the middle of the 1960s David Russell published a paper on exactly this problem. An early version of Russel's paper can be found in a 1965 Mathematics Research Center (Madison, Wisconsin) report. Russell pointed out the necessity of extending the domain to \(0< x\leq 2\ell\) in looking for the optimal control. In 1988 J. L. Lions studied boundary controls for the generalized solutions of this one-dimensional wave equation. This was followed by the studies of Il'in and the author which considered the possibility of various boundary conditions at both ends of the vibrating string, with the stipulation that specific elastic conditions, such as \(u_x(\ell, t)+ hu(\ell,t)= 0\), \(h> 0\), are applied at the point \(x=\ell\). Four functions \(\varphi(x)\), \(\varphi_1(x)\), \(\psi(x)\), \(\psi_1(x)\), are supposed to satisfy such conditions at \(t= 0\) and \(t= T\). Several types of mixed conditions are listed. Taking differences between different solutions and showing that they are identically equal to zero, the author shows that at most one solution does exist. The author then produces a function which looks like a rather strange version of the Laplace transform, which enables him to decide whether there exists a unique solution to the initial control problem. Very ingenious and lengthy manipulations in which all four of the \(\varphi(x)\) and \(\psi(x)\) functions are extended to the interval \([-2\ell, +4\ell]\) lead to explicit formulas for the required boundary control. Comparing the recent papers of this author and papers of Professor Il'in with the previous well-known works of Russell, Lions, Lasiecka, Triggiani, etc.\dots we are comparing articles which go into the nitty-gritty details of constructing specific analytical (not analytic!) solutions, reminiscent of the old paintings, where every detail is carefully taken care of, as compared with the broad brush of an impressionist (i.e. a functional analyst), who gives you an overall view of the subject of the painting, but omits the detailed insight.