Development of the fuzzy sets theory: weak operations and extension principles

The paper considers the problems that arise when using the theory of fuzzy sets to solve applied problems. Unlike stochastic methods, which are based on statistical data, fuzzy set theory methods make sense to apply when statistical data are not available. In these cases, algorithms should be based...

Full description

Saved in:
Bibliographic Details
Published inҚарағанды университетінің хабаршысы. Математика сериясы Vol. 111; no. 3; pp. 59 - 71
Main Authors S. Katsyv, V. Kukharchuk, N. Kondratenko, V. Kucheruk, P. Kulakov, D. Karabekova
Format Journal Article
LanguageEnglish
Published Academician Ye.A. Buketov Karaganda University 01.09.2023
Online AccessGet full text

Cover

Loading…
More Information
Summary:The paper considers the problems that arise when using the theory of fuzzy sets to solve applied problems. Unlike stochastic methods, which are based on statistical data, fuzzy set theory methods make sense to apply when statistical data are not available. In these cases, algorithms should be based on membership functions formed by experts who are specialists in this field of knowledge. Ideally, complete information about membership functions is required, but this is an impractical procedure. More often than not, even the most experienced expert can determine only their carriers or separate sets of the α-cuts for unknown fuzzy parameters of the system. Building complete membership functions of unknown fuzzy parameters on this basis is risky and unreliable. Therefore, the paper proposes an extension of the fuzzy sets theory axiomatics in order to introduce non-traditional (less demanding on the completeness of data on membership functions) extension principles and operations on fuzzy sets. The so-called α-weak operations on fuzzy sets are proposed, which are based on the use of separate sets of the α-cuts. It is also shown that all classical theorems of Cantor sets theory apply in the extended axiomatic theory. New extension principles of generalization have been introduced, which allow solving problems in conditions of significant uncertainty of information.
ISSN:2518-7929
2663-5011
DOI:10.31489/2023M3/59-71