ESTIMATION IN ADAPTIVE OPTIMAL-CONTROL SYSTEMS

This article considers the problem of optimal adaptive control of a dynamic object under the influence of random noise. It is assumed that there is a known control law that is optimal in the sense of a given functional in the absence of noise. An algorithm is constructed for estimation; this algorithm provides a first-approximation minimum for the deterioration in the functional. The solution that is obtained is, in a general sense, an alternative to the principle of separation for the problems of evaluation and control, and for a linear object and quadratic functional, it reduces to a Kalman-Bucy filter. An example is given.