Prognostication and minimization of heteroscedastic processes dispersion using models with multirate discretization

Abstract

Theoretical principles of designing multirate discrete systems for prognostication and minimization of the variables of maximal conditional dispersions are considered for output coordinates of one and multidimensional processes under discretization of input disturbances with small periods of sampling and output coordinates and control signals with large periods of sampling. The dynamics of processes is represented by the models of autoregression and a sliding mean as well as of autoregression and a sliding mean with additional input control signal of multirate discritization.