The maximum principle in regular optimal control problems with phase trajectory endpoints lying at the boundary of the phase constraint (Q1105833)

The authors consider an optimal control problem for a system of ordinary differential equations with regular constraints on phase and control. A necessary condition for a strong minimum is known in the form of the maximum principle, formulated in terms of a measure and a discontinuous vector function of conjugate variables. However, in problems where the optimal trajectory endpoints come out at the boundary of the phase constraint, this maximum principle degenerates into a pulse concentrated at the final instants and gives no information. Subsequent work led to an augmented maximum principle with components continuous at the final instants, the C-maximum principle, established however under very rigid a priori assumptions related to the optimal trajectory which hamper practical application. This paper continues the development of the C- maximum principle by giving minimal assumptions for the principle, which are convenient in application, one of which is a matching assumption on the constraints. The need for these assumptions is demonstrated in two examples.