ИСТИНА

The linear impulse control problem is studied via dynamic programming techniques. It is approached by introducing a hard bound on control, with a diameter of this bound tending to innity. A new denition of a feedback strategy for such problem is proposed, which takes into account the fact that in actual systems control values are large but still bounded, thus justifying the use of additional hard bound. A special attention is given to solving two-dimensional problems. For such problems the structure and properties of optimal control are studied, and optimal feedback strategy is constructed.