Кафедра системотехніки (СТ)

Постійний URI для цієї колекції

https://openarchive.nure.ua/handle/document/59

Перегляд

Перегляд Кафедра системотехніки (СТ) за автором "Berezovskyi, G."

Зараз показано 1 - 1 з 1
  • Зображення мініатюри
    Публікація
    Estimating the properties of technological systems based on fuzzy sets
    (Innovative technologies and scientific solutions for industries, 2017) Beskorovainyi, V.; Berezovskyi, G.
    The subject of the research in the article is the process of evaluating the properties of technological systems at the stages of their design and reengineering. The goal – to improve the efficiency of procedures for multi-criteria evaluation of options for constructing technological systems using the apparatus of fuzzy sets. Objectives: to search for new or modification of known functions of belonging to fuzzy sets "the best variant of building a technological system" by particular criteria in the direction of reducing the complexity of procedures for calculating their values; perform a comparative analysis of the temporal complexity and accuracy of approximation of the preferences of the decision maker with the help of monotonic membership functions; give recommendations on the practical use of monotonous membership functions in decision support systems. Common scientific methods are used, such as: decision making, utility theory, fuzzy sets and identification. The following results are obtained. The article presents the model of preferences of the decision-maker developed by the authors for evaluating individual properties of technological systems using the membership function of fuzzy sets, which allows to realize both linear and nonlinear (convex, concave, S-shaped and Z-shaped) criteria. The carried out experimental research has revealed its advantages in terms of accuracy and time complexity in comparison with the functions of Gauss, Harrington, logistic function, gluing of power functions and their modifications. Methods are proposed that reduce the time complexity of procedures for calculating the values of membership functions. Conclusions. As a result of the analysis of known membership functions for fuzzy sets, it has been established that they do not adequately reflect the preferences of the decision-maker for the characteristics of systems close to extreme values and have a relatively high computational complexity. The proposed membership function and its calculation method make it possible to increase the adequacy of multifactorial estimation models and significantly reduce the time complexity of procedures for calculating its values. Practical use of the proposed membership function in the support systems for the adoption of design and management solutions will make it possible to obtain solutions of the problems of multifactor estimation and choice of a much larger dimension or with less expenditure of computing resources practically without loss of accuracy.