Local Minimum Principle for an Optimal Control Problem with a Nonregular Mixed Constraint

Abstract

We consider the simplest optimal control problem with one nonregular mixed inequality constraint, i.e. when its gradient in the control can vanish on the zero surface. Using the Dubovitskii--Milyutin theorem on the approximate separation of convex cones, we prove a first or der necessary condition for a weak minimum in the form of the so-called local minimum principle, which is formulated in terms of functions of bounded variation, integrable functions, and Lebesgue--Stieltjes measures, and does not use functionals on the space of measurable bounded functions. Two illustrative examples are given. The work is based on results by Milyutin.