On feedback maximum principle for nonsmooth optimal impulsive control problems

摘要

We study a particular class of nonsmooth rightpoint-constrained optimal control problems being an ordinary-differential equivalent of impulsive control problems with states of bounded variation, described by a measure differential equation. The main result is a nonlocal necessary optimality condition formulated in terms of the nonsmooth Maximum Principle (NMP). At the same time, the developed optimality condition defines an algorithm for discarding local extrema of NMP.