Higher-order convex approximations of Young measures in optimal control (Q1871991)

Nonconvex optimization problems arising in different fields of variational calculus, optimal control or noncooperative games theories may not possess a classical solution because approximate solutions show typically rapid oscillations. A solution of such problems is often constructed by means of Young measures. In this paper the general theory of convex approximation scheme of Young measures is presented and its applications to optimal control problems with an explicit constraint on controls are considered. A suitable linear operator approximating the test integrands is constructed. The adjoint operator gives an approximation of a generalized Young measure, considered as a linear functional on test integrands. An adjoint-operator approach, combined with quasi-interpolation of the test integrands, is systematically used. The applicability of the results is demonstrated on an optimal control problem for an elliptic system. In the end of the paper the results of numerical calculations for series of illustrative examples are given.