The normal numbers of the fuzzy systems and their classes

Abstract

R-fuzzy set is defined in this paper, which is regarded as the generalization of the Zadeh fuzzy set. By means of CRI method, some fuzzy systems are constructed by suitably using several kinds of R-fuzzy sets as fuzzy inference antecedents, such as interpolation fuzzy system, Bernstein fuzzy system, Lagrange fuzzy system and Hermite fuzzy system. A notion of the normal number of the fuzzy system is defined here, we have shown that all fuzzy systems are able to be classified as three classes such as the normal fuzzy systems, the regular fuzzy systems and the singular fuzzy systems under the significance of the normal numbers of fuzzy systems. Finally, the generalized Bernstein polynomial is obtained by constructing Bernstein fuzzy system, it is proved that the generalized Bernstein polynomial is uniformly convergent in C[a, b] under a weaker condition, and it is pointed out that there exist generalized Bernstein polynomials to be not convergent in C [a, b] by use of constructing a counterexample.KeywordsR-fuzzy set, fuzzy system, the normal number of fuzzy system, universal approximation property for fuzzy system CitationLi H X, Yuan X H, Wang J Y, et al. The normal numbers of the fuzzy systems and their classes. Sci

Approximation properties of ELM-fuzzy systems for smooth functions and their derivatives

“…Definition 2.2 (Mo and Liu [17]). Suppose that f A C½0; 1, then the generalized Bernstein polynomial of degree nðn Z 1Þ is defined as…”

“…Different from generalized Bernstein fuzzy system (3), β k ðk ¼ 0; …; nÞ in ELMBernstein fuzzy system (17) are unknown parameters which need to be obtained by ELM method and some given samples.…”

“…The ELMBernstein fuzzy system is chosen as Eq. (17). The corresponding result is shown in Table 3, where the first component denotes the approximation error about f 2 , the second component denotes the approximation error about f 0 2 and the third component denotes the approximation error about f″ 2 .…”

“…In [14,15], formulae which can determine the minimal upper bounds on the number of fuzzy rules for TS fuzzy systems are established. Besides, the approximation properties of other types of fuzzy systems, such as Boolean fuzzy systems [16] and generalized Bernstein fuzzy systems [17], are investigated. In real world applications, many models, such as control system and decision model, operate in dynamic environment.…”

Grey qualitative modeling and control method for subjective uncertain systems

“…Thus, fuzzy logic [1−6] , qualitative simulation [7−11] , grey system theory [12,13] etc., are proposed to solve the problem. Fuzzy logic based methods are interested in establishing fuzzy relationships between system states x and y [14,15] :…”