On optimality conditions for optimal control problem in coefficients for \(\Delta_p\)-Laplacian (Q472260)

The considered optimal control problem involves the minimization of a functional describing the discrepancy between the solution \(y\) of a nonlinear Dirichlet boundary value problem for the equation in the \(p\)-Laplacian \(\Delta_p(u,y)=f\) and \(y_d\), which is a given distribution. In this equation, \(u\) means the control and belongs to a class of admissible controls which turns out to be a convex set with an empty topological interior and \(f\) is a given distribution. In the Preliminaries, the existence of a weak solution is shown and the set of admissible pairs \((u, y)\) is defined. Further, the existence of at least one solution to the stated optimal control problem is proved. Since for a given distribution \(f\), the mapping \(u\to y(u)\) is not Fréchet-differentiable on the class of admissible controls and this class has an empty topological interior, the classical approach to derive optimality conditions will not be applied. Instead, one makes use of the notion of a quasi-adjoint state to an optimal solution, introduced by \textit{S. Y. Serovajskiy} in [``Variational inequalities in non-linear optimal control problems'', Methods Means Math. Model 1, 156--169 (1977)]. In this way, the Lagrange functional associated to the optimal control problem is introduced and a set of preliminary results concerning its differentiability is shown. The next sections deal with the presentation of a formal approach to derive first order optimality conditions, introduction of the quasi-adjoint state, the study of the well-posedness of the obtained variational problem and the asymptotic behavior of its solutions. Finally, in the last section, first order optimality conditions for the stated optimal control problem are derived.