Guaranteed control design under $L_p$-compact constraints on the disturbance

Abstract

The paper deals with the problem of optimization of a guaranteed (worst case) result for a control system described by an ordinary differential equation. The disturbances as functions of time are subject to functional constraints belonging to a given family of constraints. The latter family is known to the side that forms the control actions. The controlling side uses positional full-memory strategies and does not observe the disturbance. When the constraints family consists of $L_p$-compact sets the optimal guaranteed result is non-improvable in the sense that it coincides with that obtained in the class of quasi-strategies -- nonanticipatory transformations of disturbances into controls. In this paper for the effectiveness of implemented control algorithm an additional condition on the system and appropriate ways of constructing an optimal strategy are specified.