Approximate aggregation of a class of nonstationary linear stochastic systems (Q1914132)

Many real systems are described by linear stochastic models with a block-triangular dynamics matrix. The paper under review deals with the approximation of a high-order model by a model of lower order. The parameters of the approximate system are defined so that the variance of the error of approximation of the essential components is minimized, while the structure of the main subsystem and the desired part of the poles of the auxiliary subsystem are retained. A numerical example is given to illustrate the efficiency of the method.