Zubretskaya N. Development of scientific fundamentals for the multi-criteria evaluation and forecasting of the industrial product quality.

The thesis deals with important scientific and applied problem solution of reliability assurance of multi-criteria evaluation and industrial products quality forecasting to the stages of design, construction and operation by using adaptive methods and intelligent decision support systems. Propose a conceptual model of information system of multi-criteria evaluation, forecasting and quality control of industrial products. The method of reliability increase of products composite quality index based on simultaneous use of desirability function and fuzzy membership functions of sigmoid and gauss types is improved. The system of neural-fuzzy accuracy control and technological processes stability which allows providing the normalized value of products quality characteristic under control is developed. The neural network methods for obtaining of multi-criteria relation between structural specification of products and parameters of technological processes have been established to provide to reduce the uncertainty of making decisions for identification of rational parameters during design and manufacturing of industrial products. The adaptive methods and models for obtaining reliable forecasts of quality level and competitiveness of industrial products are developed.