Drozdova T. Method of quality assessment of hierarchical systems with the use of fuzzy sets

The hierarchical structure of systems from the point of view of the general theory of systems is investigated. The analysis of hierarchy of quality assurance system of complex system is carried out, its vertical decomposition is considered. The main tasks of the studied system for each level of its hierarchy are determined and the requirements to it are formed. Developed "quality observer" as a situational system with fuzzy logic in solving qualimetric coordination, which is responsible for self-monitoring and issuance of recommendations for the restoration of qualimetric characteristics of the system in conditions of uncertainty, dynamism and instability of characteristics over time. The main stages of ensuring the compatibility of the "quality observer" in the monitoring system are identified. The definition of the main functions of the "quality observer" at the first and second functional levels of the system, the choice and justification of the mathematical apparatus for its construction - fuzzy-multiple method. The number of terms required for the description and the number of reference situations, the type of membership function are determined. The model of the block of fuzzification of input statistical data on quality of object of research taking into account both "systematic", and "random" components of uncertainties of expert estimation is considered. A study of the quality monitoring system as a queuing system that drifts over time, and the development of a method of operation of the proofreader, which is based on its work has expert data. The use of "universal scales" for quality measurement is proposed, the advantage of which is their relative independence from the operating conditions of the fuzzy system. The adequacy of the proposed method of quality monitoring is experimentally proved. The object of approbation is the procedure for assessing the quality of the educational process. Computer modeling of fuzzy expert system is carried out.