An application of the duality theory of Toland to a problem of optimal control (Q2639335): Difference between revisions

The paper describes the application of \textit{J. F. Toland}'s duality theory [see ibid. 66, 399-415 (1978; Zbl 0403.90066); Arch. Ration. Mech. Anal. 71, 41-61 (1979; Zbl 0411.49012)] to a problem of optimal control. The cost function can be written as the difference of two convex functions. The constraints appear as an affine-linear operator equation of state with a nonempty, convex and closed set as restriction of control. Through a suitable construction one can use Toland's dual problem to obtain conditions for optimality and by certain assumptions the Fréchet-differentiability of the dual cost function on the whole space. The application of the results to a quadratic non-convex problem of optimal control for a homogeneous Dirichlet problem with an elliptic differential equation is given.