Melnyk Y. Methodology for increasing control quality indexes and telecom networks conditions monitoring with the use of fuzzy set theory

The dissertation has scientific problem solution of the development methodology of increasing control quality indexes and telecoms networks condition monitoring with the use of fuzzy set theory. This is a topical problem and it has important scientific and practical value. Improved a fuzzy forecasting method of network device condition on the ground of a fuzzy generic parameter with the use of fuzzy multinominal coefficients, which serves for forecasting subsystems enhancement and distributed control systems and an object management condition monitoring TKM. The method of control telecommunication network on the base of taking into account the summarized condition index of the operator TKM has been developed. This method allows to make condition evaluation of multidimensional objects with linguistic variables and to use for researching. The method of neural network controller structure and non-stationary object control has been developed. This method proposes the combined use of neural networks and statistical approaches, when the statistical component provides revealing of a significant change in the object’s characteristics and the launch of the learning mechanism. The proposed approach reduces response time to changes of the control object and diminishes mean-root-square error faster than traditional methods. The control model of channel provision has been developed as well on the ground of direct fuzzy production output rule and routing algorithm for the "Manager-Agent" model, which makes properly routes matrix when appears intensive network loads and interfering factors of different nature’s origin.