No label defined (Q4486985)

The paper considers an optimal control problem for the elliptic equation \[ \text{div}(A(x)\nabla y)= f(y, u)\quad\text{in }\Omega, \Omega\subset \mathbb{R}^n, \] with the zero Neumann boundary conditions on \(\partial\Omega\) and additional integral constraints on the state \(y\). The admissible controls \(u\) belong to a bounded convex set \(U\subset L_\infty(\Omega)\) and it is assumed that \(f_y(y, u)\geq\mu> 0\) for all \((y,u)\in \mathbb{R}^2\). The authors derive second-order necessary optimality conditions expressed in the form of nonnegativity of the second derivative of the Lagrangian on a special set of pairs \((h, y_h)\in L_\infty(\Omega)\times W^1_p(\Omega)\) satisfying the linearized state equation.