Optimal Stabilization Problem for the Quasilinear System with Controllable Parameters

Abstract

The main concern of this paper is the problem of optimal stabilization of a quasilinear stochastic system with controllable parameters. Systems of this type are described by linear stochastic differential equations with multiplicative noises whose matrices, in general case, are nonlinear functions of control. The performance criterion is a modification of the classic quadratic performance cost. The goal is to minimize the criterion on the set of admissible control processes. This formulation of the problem is interesting because it allows us to study a wide range of optimization problems of linear systems with multiplicative perturbations, including: optimization of design parameters of the system, the problem of optimal stabilization under constraints on the gain matrix of the linear regulator in the form of inequalities, the problem of optimal stabilization of linear stochastic systems under information constraints. The main result of this paper is the necessary conditions for the optimal vector in the problem of stabilization of a quasilinear stochastic system with controllable parameters.The numerical gradient-type procedure for synthesis of the optimal stabilizing vector is also proposed. In addition, using obtained results we construct the algorithm for synthesis of a suboptimal time-dependent control. The result of the proposed algorithm is piecewise constant control, which gives the value of the criterion is guaranteed not worse than for the optimal stabilizing vector. This algorithm is relatively simple and one may use it for calculations in real time. The obtained results are applied to the problem of optimal stabilization under information constraints, in which the necessary optimality conditions are also obtained and the gradient-type procedure for the synthesis of the optimal control is proposed. The use of the obtained results is demonstrated by a model example.