![]() | Only 14 pages are availabe for public view |
Abstract The notion of fuzzy sets was created by L.A.Zadeh [61] in 1965. This notion played an important role in applications and in such mathematical theories as Logic and Topology. In 1968, Chang [11] used this concept to introduce the concept of fuzzy topological spaces. Subsequently, Wong [57, 58], Lowen [26], and others extend some basic concepts from general topology to fuzzy subsets and developed a theory of fuzzy topological spaces. We denote by FTS the category of Lowen’s fuzzy topological spaces and their continuous functions. In 1982, Lowen [30], introduced the concept of fuzzy neighbourhood spaces the category of which is denoted by FNS. Many papers have appeared analyzing and making use of this concept. Some papers showed the distinguished internal characterizations of FNS (e.g. [42], [59]). Given a lower semi-continuous triangular norm T, Morsi [19, 46, 47] introduced the two categories of fuzzy T-Iocality spaces and of fuzzy T-neighbourhood spaces, denoted by T-FLS and T-FNS, respectively. In the special case T = Min, the categories T-FLS and T-FNS coincide with the categories of fuzzy T-neighbourhood -base spaces, denoted by FNBS [44] and FNS, respectively. On another side, ill 1982, Z. Pawlak introduced the concept of rough sets and of approximation spaces [50] in accordance of an indescernibility relation (e.g. equivalence relation). This concept played an important role in applications especially in medicine and artificial intelligence (e.g.[5I]). In 1985, Z. Pawlak introduced the relationship between rough sets and fuzzy sets [52]. In 1986, L. Farinas and H. Prade introduced the concept of fuzzy rough sets and of fuzzy approximation spaces [17]. We now give a brief overview of the contents of the thesis. It consists of five chapters: Chapter I is devoted to give an exposition of some needed definitions and results to be used throughout the thesis. In Chapter II, we introduce the axioms of fuzzy T-upper (lower) approximation operators on IX for a lower semi-continuous triangular norm T and study the relationships between them and that between each of them and T-similarity relation on X. We generalize the definition of fuzzy upper approximation operators introduced by L. Farinas and H. Prarle in [17]. In the case T = Min, the reader may notice that our definition of the upper approximation operator AR coincides on the corresponding definition introduced in [17]. We prove many properties of A more than that mentioned in [17];·whence we extend the theory of fuzzy rough sets. We introduce a definition of fuzzy T-Iower approximation operator of fuzzy rough set (X, .R) differs that introduced in [17]; in the case T = Min, the reader may notice that our definition of the lower approximation operator &is different from the corresponding definition introduced in [17]. Their definition does not satisfy our identities AA~ = Afl and AAfl = Afl for all ~EIx which satisfied in the classical case in [50]. On the other hand, their definition is designed to satisfy A(l-fl) = I-A~ for all flEIx. This identity is not valid for our definition. |