Optimal control problems with weakly converging input operators (Q1355012): Difference between revisions

The variational convergence of a sequence of optimal control problems \[ \min \{J(u,y) : A_hy = B_hu,\;(u,y) \in U \times Y \} \tag \(P_h\) \] is studied, where \(u\) and \(y\) denote, respectively, the control and the state of the system belonging to some Banach spaces. More precisely, the authors are looking for conditions under which there exists a limit problem and try to identify its form. In the problems \((P_h)\) the cost functionals could have, for instance, the following integral form \(\int_{\Omega}f_h(x,y(x),u(x))dx\). As concerns the state equations the operators \(A_h\) are supposed to \(G\) converge to the operator \(A\), while for the input operators \(B_h\) the following three cases are considered: (i) \(B_h\) converge continuously to \(B\), (ii) \(B_h\) are linear but may converge to \(B\) only weakly, (iii) \(B_h\) may be nonlinear, weakly converging, but local, i.e., \(B_h(u) = B_h(x,u(x))\). The obtained results show that in the case (i) the limit problem \((P_{\infty})\) is of the same type as the problems \((P_h)\). On the other hand, in cases (ii) and (iii) the limit state equation may disappear but then an extra penalizing term must be added in the limit cost. Such terms are explicitely characterized in the paper. Some examples illustrating the results are also presented.