Gorobets' O. Control objects structural identification method, which are described by the simultaneous equation systems

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0401U002005

Applicant for

Specialization

  • 05.13.03 - Системи та процеси керування

21-06-2001

Specialized Academic Board

Д50.052.02

Essay

The structural identification criteria offered in this thesis are at the base of the developed method devoted to the accuracy prediction enlargement of the control objects parameters in the conditions of the limited sample. The objective of this method is the selection of the optimal structure of simultaneous equation system from the given model set by one of the selection criteria. The statistical properties of these criteria are investigated by Monte-Carlo simulations. Analytical expres-sions of the mean square error of prediction matrixes are deduced for both structural and reduced forms of systems. The biased, true and overfitted classes of models of simultaneous equation systems are considered. Theoretical results of the thesis are inculcated on the regional level. Application sphere of the developed method is the control systems of the multi related, stochastic processes.

Similar theses