A fuzzy difference equation with an application

The paper deals with a first order difference equation in fuzzy setting [cf. \textit{E. Y. Deeba, A. De Korvin} and \textit{E. L. Koh}, Differ. Equ. Dyn. Syst. 4, No. 2, 149-156 (1996; Zbl 0873.39002)]. It is proved that the solutions \(x_n=g_n(x_0)\) of the recurrence equation \(x_{n+1}=f(x_n)\) with continuous increasing functions \(f\) and \(g\) can be extended to fuzzy numbers [cf. \textit{D. Dubois} and \textit{H. Prade}, Int. J. Syst. Sci. 9, 613-626 (1978; Zbl 0383.94045)]. However the final formula is proved in the case of endpoints of the considered level intervals only. The fuzzified formula is false even in the case of the linear recurrence with constant coefficients: \(x_{n+1}=wx_n+q\). E.g., if \(w=[a,b]\) is an interval coefficient with \(0<a<b\), then in the formula from Theorem 4.1 it is suggested that \(u=\frac{1-w}{1-w}=1\) (case \(n=1\)). But in interval arithmetic we have the following three cases: a) if \(1\in[a,b]\), then \(n\) is indefinite (division by zero); b) if \(b<1\), then \(u=[\frac{1-b}{1-a},\frac{1-a}{1-b}]\neq 1\) (interval); c) if \(1<a\), then \(u=[\frac{a-1}{b-1},\frac{b-1}{a-1}]\neq 1\) (interval). Moreover, the used notion of fuzzy numbers is not a useful generalization of crisp numbers because of the continuity of membership functions.

Recommendations

Cites work

Cited in

(30)

This page was built for publication: A fuzzy difference equation with an application

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4352513)