On a finite element method for measure-valued optimal control problems governed by the 1D generalized wave equation (Q1747409)

The summary of this paper is clearly and comprehensibly described by the authors: ``An optimal control problems governed by the initial-boundary value problem for the 1D wave equation having \textit{variable coefficients} and with the control spaces of either measure-valued functions or vector measures are considered. Bilinear finite element discretizations for the state and adjoint state equations by the regularized Galerkin method are constructed. They are different from those applied in [\textit{K. Kunisch} et al., SIAM J. Control Optim. 54, No. 3, 1212--1244 (2016; Zbl 1343.35239)] for the case of constant coefficients. Stability bounds and error estimates in several norms are studied. Finally, error estimates for both the optimal state variable and the cost functional are presented. This paper is a short version of the paper [\textit{P. Trautmann} et al., Math. Control Relat. Fields 8 (2), accepted for publication (2018)], where full proofs of all the result presented here can be found.'' The paper is very well written and organized. After an introduction, three main sections follow. In the first section, the initial-boundary value problem for the 1D generalized wave equation considered as the state equation is defined. The bounds for its weak solution are introduced. Then the control problem is formulated and a bound for its solution is presented. Finally, the weak solution to the adjoint problem for the state equation is defined. Relations between the optimal control and corresponding optimal adjoint state are shown. The second section is devoted to discretization. Stability bounds and error estimates in several norms for the discrete state equation, for a solution to the discrete optimal control problem, and for the discrete adjoint state equation are given. In this section, the time-stepping formulation of the discrete state and adjoint state equations are also presented. The last section contains final results of the paper -- the error estimate for the optimal discrete state variable (whose orders depend on the choice of the control space and the data smoothness) and the error estimate for the cost functional of the discrete optimal control.