Open AccessALGORITHMS OF SYSTEM SYNTHESIS FOR STABILIZING UNCERTAIN DYNAMIC OBJECTS
Author(s) -
Uktam Mamirov
Publication year - 2020
Publication title -
technical science and innovation/tehnika fanlari va innovaciâ
Language(s) - English
Resource type - Journals
eISSN - 2181-1180
pISSN - 2181-0400
DOI - 10.51346/tstu-02.21.1-77-0011
Subject(s) - stability (learning theory) , object (grammar) , constructive , inverse , decomposition , matrix (chemical analysis) , mathematical optimization , computer science , basis (linear algebra) , algorithm , mathematics , artificial intelligence , ecology , materials science , geometry , process (computing) , machine learning , composite material , biology , operating system
The article deals with the formation of stable algorithms for the system synthesis for the stabilizing uncertain dynamic objects based on the method of local optimization in the presence of approximate mathematical models. The article analyzes the issues of building adaptive control systems using the concept of roughness, taking into account the assessment of the maximum allowable discrepancy between the object and its model. Some of the most constructive algorithms for determining pseudo-inverse matrices are given. When calculating the pseudo-inverse matrix of the control object, modified QR decomposition algorithms are used, obtained by deleting or assigning a column. The obtained algorithms allow us to conclude that the stabilization systems can be built on the basis of the local optimization method in the presence of approximate mathematical models. At the same time, it turns out that the asymptotic stability, i.e., the limited output with a limited input, can be achieved with sufficiently rough estimates of the object parameters and external perturbations using regular methods.