Algorithm for real-time optimization of incompletely determined linear control systems. II. Constantly acting perturbations (Q1324053)

In the first part of this article [Autom. Remote Control 54, No. 4, pt. 1, 555-563 (1993); translation from Avtom. Telemekh. 1993, No. 4, 34-43 (1993; Zbl 0813.49028)], we investigated the problem of guaranteed optimization of a control system with indefiniteness in the initial state. In practice, it is typical to have, in additional to perturbations in the initial state, perturbations that act on the system throughout the period of its operation. Probabilistic models of constantly acting perturbations have been studied in depth within the framework of stochastic optimal control and a theory of optimal control with allowance for multivariate perturbations began to be developed (see the References). In this paper, we construct algorithms for the operation of estimators and a controller that provides optimal control in real time for a linear dynamic system under the influence of constantly acting perturbations.