Exact controllability of wave equations with internal control in moving boundary domain (Q6569355)

This article discusses the distributed control of the \(1\)-d wave equation set in a domain of the form \(\{(t,y) \text{ such that } y \in (0, a_k(t)) \}\), with \(a_k(t) = 1+ k t\), and \(k \in (0,1)\), when the control domain also is a moving domain, of the form \(\{ (t,y) \text{ such that } y \in (m a_k(t), n a_k(t)) \}\), for some \((m,n) \subset (0,1)\). This allows to perform a change of variable to recast the problem in a cylindrical domain \((0,T) \times (0,1)\) with control in the interval \((m,n)\), the drawback being that the equation then has time and space dependent coefficients. This wave equation is then analyzed using a multiplier argument, i.e. multiplying the equation by \(p \partial_x z\), for some \(p\) suitably chosen (and linear outside of \((m,n)\)), where \(z\) is the solution of the equation in the cylindrical domain. The time given by this approach does not seem optimal, and using characteristics could possibly give other insights.