Elsevier

Information Sciences

Volume 248, 1 November 2013, Pages 31-49
Information Sciences

Fuzzy rough sets based on generalized residuated lattices

https://doi.org/10.1016/j.ins.2013.03.051Get rights and content

Abstract

This paper is devoted to propose generalized L-fuzzy rough sets as a further generalization of the notion of L-fuzzy rough sets. A quadruple of approximation operators are defined to suit the situation when generalized residuated lattices are non-commutative. Generalized L-fuzzy rough sets are characterized from both constructive and axiomatic approaches. In the constructive approach, various classes of generalized L-fuzzy rough sets are investigated. Moreover, the relationship between generalized L-fuzzy rough sets and L-topologies on an arbitrary universe is discussed with generalized lower and upper sets. As an application of generalized L-fuzzy rough sets, fuzzy rough sets are proposed and studied on the unit interval [0, 1], which are based on generalized residuated lattices induced by left-continuous pseudo-t-norms.

Introduction

The origins of residuation theory lie in the study of ideal lattices of rings in the 1930s. Ward and Dilworth investigated residuated lattices in a series of important papers [8], [9], [47], [48], [49]. Since that time, there have been substantial researches regarding some specific classes of residuated structures, such as [4], [5], [11], [13], [19], [20]. BL-algebras [19] are the generalization of the residuated structures on the unit interval [0, 1] induced by continuous t-norms. The monoidal t-norms based logic (MTL, for short) was obtained in [11] by considering left-continuous t-norms. When we consider the conjunction to be non-commutative, the corresponding truth structures are weak-pseudo-BL-algebras [13] or pseudo-MTL-algebras [20]. Similarly to BL-algebras and MTL-algebras, weak-pseudo-BL-algebras or pseudo-MTL-algebras are the generalization of the generalized residuated lattices on [0, 1] induced by left-continuous pseudo-t-norms [13], [17].

Rough sets theory was proposed by Pawlak [33], [34] as a formal tool to deal with imprecision and uncertainty in data analysis. Usefulness of rough sets theory has been fully demonstrated by its applications [42], [43].

In recent years, two different approaches have been formed to investigate rough sets theory, i.e., the constructive way and axiomatic way. In the constructive approach, lower and upper approximation operators are constructed and discussed from a binary relation. Diverse forms of rough sets have been proposed and studied with different binary relations [10], [12], [23], [25], [28], [32], [33], [38], [39], [55], [58], [62], [63]. Moreover, some of them are investigated with addition topological or algebraic structures [18], [22], [29], [35], [37], [53]. In contrast to the constructive approach, abstract lower and upper approximation operators are characterized by certain axioms [6], [26], [30], [31], [41], [44], [50], [54] in the axiomatic approach.

Morsi and Yakout studied fuzzy rough sets based on a left-continuous t-norm and its residual implication in [31]. In [38], Radzikowska and Kerre defined a broad family of (I,T)-fuzzy rough sets which is determined by arbitrary implication I and any t-norm T, where the work of Morsi and Yakout is a special case of (I,T)-fuzzy rough sets. Wu et al. discussed (I,T)-fuzzy rough sets from the axiomatic approach in [51]. However, (I,T)-fuzzy rough sets still investigate fuzzy rough sets on the unit interval. Since a left-continuous t-norm and its residual implication form an MTL-algebra on [0, 1], Radzikowska and Kerre proposed fuzzy rough sets based on residuated lattices in [39], [40], which were also called L-fuzzy rough sets for short and further studied in [52]. The axiomatic studies for L-fuzzy rough sets were made by She and Wang [41], in which residuated lattices are assumed to be regular. Hao and Li discussed the relationship between L-fuzzy rough sets and L-topologies in [18]. As a further discussion on [18], [41], Ma and Hu investigated topological structures of L-fuzzy rough sets and the relationship between upper (resp. lower) sets and lower (resp. upper) L-fuzzy rough approximation operators in [29]. Although Radzikowska and Kerre defined L-fuzzy rough sets based on generalized residuated lattices in [39], only left implication was considered. Most of properties of L-fuzzy rough sets discussed in [18], [29], [39], [40], [41], [52] are still limited to the commutative and integral generalized residuated lattices (residuated lattices, for short). Thus fuzzy rough sets based on generalized residuated lattices (generalized L-fuzzy rough sets, for short) should be studied. The aim of the present paper is to properly propose generalized L-fuzzy rough sets and then to investigate them from both constructive and axiomatic approaches, where the conjunctions are not always commutative.

This paper gives a complete study on the generalized L-fuzzy rough sets. In Section 2, we recall some fundamental concepts and related properties, especially the properties of generalized residuated lattices with examples to show the differences between generalized residuated lattices and residuated lattices. In Section 3, we propose generalized L-fuzzy rough sets, and investigate their properties and several classes of generalized L-fuzzy rough sets. Section 4 characterizes generalized L-fuzzy rough sets from the axiomatic approach. In Section 5, we discuss the relationship between generalized L-fuzzy rough sets and L-topologies. We construct a one-to-one correspondence between the set of L-fuzzy preorders and the set of left (resp. right) Alexandrov L-topologies on an arbitrary universe. In Section 6, we propose and study generalized I-fuzzy rough sets, which are based on generalized residuated lattices induced by left-continuous pseudo-t-norms. When the fuzzy relation is an I-fuzzy equivalence relation, generalized I-fuzzy rough sets are proved to be complete completely distributive (CCD, for short) rough sets [7]. In the final section, our researches are concluded.

Section snippets

Preliminaries

Definition 2.1

A generalized residuated lattice is an algebra (L, ∧, ∨, ⊗, →, ⇝, 0, 1, ⊤) such that the following conditions hold.

  • (GR1)

    (L, ∧, ∨, 0, 1) is a bounded lattice with the least element 0 and the greatest element 1.

  • (GR2)

    (L, ⊗, ⊤) is a monoid.

  • (GR3)

    For all a, b, c  L we have the equivalence:abciffabciffbac.→and ⇝ are called the left and right implication of ⊗, respectively.

Definition 2.2

A generalized residuated lattice (L, ∧, ∨, ⊗, →, ⇝, 0, 1, ⊤) is said to be

  • (1)

    commutative (resp. non-commutative) if ⊗ is commutative (resp. non-commutative),

  • (2)

    integral

Generalized L-fuzzy rough sets

In this section, we propose generalized L-fuzzy rough sets and investigate their properties and several classes of generalized L-fuzzy rough sets. When generalized residuated lattices are non-commutative, we have two implications, but not one. As a consequence, a generalized L-fuzzy rough set is not a pair, but a quadruple of approximation operators, which is in fact two pairs of approximation operators.

Assume that X is a nonempty universe and R is an arbitrary L-fuzzy relation on X, then the

Axiomatic characterizations of generalized L-fuzzy rough approximation operators

In this section, we study the axiomatic characterizations of generalized L-fuzzy rough approximation operators. It is shown that generalized L-fuzzy rough approximation operators can be characterized by abstract L-fuzzy sets operators, which guarantees the existence of certain types of L-fuzzy binary relations producing the same generalized L-fuzzy rough approximation operators.

Definition 4.1

Let a mapping ψ:LX  LX. Then

  • (1)

    ψ is called a left lower L-fuzzy rough approximation operator if it satisfies the following

Topological structures of L-fuzzy approximation space

In this section, the notions of lower and upper sets are extended and the topological structures of L-fuzzy approximation space are investigated.

In Pawlak rough set theory, both (R¯,R-1̲) and (R-1¯,R̲) are Galois connections. In [1], [16], [36], [56], the fuzzy Galois connection was also studied. In [18], [29], it has been pointed out that the pairs (R¯r,R-1̲l) and (R-1¯r,R̲l) are fuzzy Galois connections using the subsethood operator [14], when generalized residuated lattices are commutative.

Generalized I-fuzzy rough sets

In this section, we investigate generalized I-fuzzy rough sets as an application of generalized L-fuzzy rough sets, where I denotes the unit interval [0,1].

A binary operation ∗:I × I  I is called a pseudo-t-norm [13], [17] on I if it is associative, increasing in each variable and has a unit element 1. In [13], it was pointed out that all the continuous pseudo-t-norms are commutative, so there does not exist any continuous non-commutative conjunction on I. A pseudo-t-norm ∗ is left-continuous if

Conclusion

In this paper, generalized L-fuzzy rough set was proposed with a quadruple of approximation operators. Generalized L-fuzzy rough sets were investigated from both constructive and axiomatic approaches. Moreover, generalized lower and upper sets were proposed to investigate the relationship between generalized L-fuzzy rough sets and L-topologies, and the connections between L-interior (resp. L-closure) operators and generalized approximation operators. Especially, the one-to-one correspondence

Acknowledgements

The authors thank the anonymous referees and the Editor-in-Chief, Professor Witold Pedrycz for their valuable comments and suggestions in improving this paper. This research was supported by the National Nature Science Foundation of China (Grant No. 61179038).

References (63)

  • Z.M. Ma et al.

    Topological and lattice structures of L-fuzzy rough sets determined by lower and upper sets

    Information Sciences

    (2013)
  • J.S. Mi et al.

    An axiomatic characterization of a fuzzy generalization of rough sets

    Information Sciences

    (2004)
  • N.N. Morsi et al.

    Axiomatics for fuzzy rough sets

    Fuzzy Sets and Systems

    (1998)
  • Z. Pei et al.

    Topology vs generalized rough sets

    International Journal of Approximate Reasoning

    (2011)
  • K. Qin et al.

    On the topological properties of fuzzy rough sets

    Fuzzy Sets and Systems

    (2005)
  • A.M. Radzikowska et al.

    A comparative study of fuzzy rough sets

    Fuzzy Sets and Systems

    (2002)
  • Y.H. She et al.

    An axiomatic approach of fuzzy rough sets based on residuated lattices

    Computers and Mathematics with Applications

    (2009)
  • H. Thiele

    On axiomatic characterizations of crisp approximation operators

    Information Sciences

    (2000)
  • W.Z. Wu et al.

    On characterizations of (I,T)-fuzzy rough approximation operators

    Fuzzy Sets and Systems

    (2005)
  • Y.Y. Yao

    Two views of the theory of rough sets in finite universes

    International Journal of Approximation Reasoning

    (1996)
  • Y.Y. Yao

    Constructive and algebraic methods of the theory of rough sets

    Information Sciences

    (1998)
  • Y.Y. Yao

    A comparative study of fuzzy sets and rough sets

    Information Sciences

    (1998)
  • W. Zhu et al.

    Reduction and axiomization of covering generalized rough sets

    Information Sciences

    (2003)
  • Q.Y. Zhang et al.

    Continuity in quantitative domains

    Fuzzy Sets and Systems

    (2005)
  • Q.Y. Zhang et al.

    Section-retraction-pairs between fuzzy domains

    Fuzzy Sets and Systems

    (2007)
  • Q.Y. Zhang et al.

    Fuzzy complete lattices

    Fuzzy Sets and Systems

    (2009)
  • W. Zhu

    Generalized rough sets based on relations

    Information Sciences

    (2007)
  • W. Zhu

    Topological approaches to covering rough sets

    Information Sciences

    (2007)
  • R. Bělohlávek

    Fuzzy Galois connections

    Mathematical Logic Quarterly

    (1999)
  • R. Bělohlávek

    Fuzzy Relational Systems: Foundations and Principles

    (2002)
  • R. Bělohlávek et al.

    Fuzzy Equational Logic

    (2005)
  • Cited by (52)

    • Rough L-fuzzy sets: Their representation and related structures

      2022, International Journal of Approximate Reasoning
    View all citing articles on Scopus
    View full text