Next Article in Journal
Stability Analysis of a Stage-Structure Predator–Prey Model with Holling III Functional Response and Cannibalism
Next Article in Special Issue
Pseudo Overlap Functions, Fuzzy Implications and Pseudo Grouping Functions with Applications
Previous Article in Journal
The Relationship between Fuzzy Reasoning Methods Based on Intuitionistic Fuzzy Sets and Interval-Valued Fuzzy Sets
Previous Article in Special Issue
On Implicative Derivations of MTL-Algebras
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Comparison of Overlap and Grouping Functions

School of Electronics and Information Engineering, Taizhou University, Taizhou 318000, China
Axioms 2022, 11(8), 420; https://doi.org/10.3390/axioms11080420
Submission received: 17 July 2022 / Revised: 10 August 2022 / Accepted: 17 August 2022 / Published: 20 August 2022
(This article belongs to the Special Issue Non-classical Logics and Related Algebra Systems)

Abstract

:
This paper investigates the pointwise comparability of overlap and grouping functions which obtained by Bustince et al.’s and Bedregal et al.’s generator pairs, respectively. Some necessary and sufficient conditions for the comparison of these functions are proved. We also introduce some compositions of these functions and study the order preservation of these compositions.

1. Introduction

Overlap function introduced by Bustince et al. [1] is a particular type of aggregation function [2]. Its dual concept is the grouping function [3]. In recent years, those two concepts have attracted a wide range of interests. For applications, they have been successfully applied to many domains, such as image processing [1,4], classification [5,6] and decision making [7,8]. For theoretical research, general overlap and grouping functions [9,10], N-dimensional overlap functions [11], Archimedean overlap functions [12], general interval-valued overlap functions [13], complex-valued overlap and grouping functions [14,15], quasi-homogeneous overlap functions [16], pseudo-homogeneous overlap and grouping functions [17], overlap functions on bounded lattices [18], overlap and grouping functions on complete lattices [19] have been introduced. Many fuzzy concepts derived from overlap and grouping functions, such as generalized interval-valued OWA operators [20], residual implications [21,22], (G,N)-implications [23], binary relations [24], (IO, O)-fuzzy rough sets [25] and so on.
In the study of overlap and grouping functions, the study of their properties accounts for a large proportion and play an important role. Bustince et al. [1] gave an alternative characterization of overlap functions by their generator pairs. Bedregal et al. [26] gave an alternative characterization of grouping functions in a similar way. Dimuro et al. [27] introduced the additive generators of overlap and grouping functions. Qiao and Hu [28] studied the interval additive generators of interval overlap and grouping functions. They [29] also introduced the multiplicative generators of overlap and grouping functions.
We have already known that there is a partial order between two t-norms T 1 and T 2 , i.e., T 1 T 2 if T 1 ( a , b ) T 2 ( a , b ) for all ( a , b ) [ 0 , 1 ] 2 (see [30], Chapter 6). Klement et al. [31] presented a necessary and sufficient condition for the comparability of continuous Archimedean t-norms. There also exist some pointwise comparison results of fuzzy implications (see [32], Chapter 1). However, comparatively little investigation has been made on the comparability of overlap/grouping functions. Bustince et al. [1] defined the pointwise order of two overlap functions O 1 and O 2 , i.e., O 1 O 2 if O 1 ( a , b ) O 2 ( a , b ) for all ( a , b ) [ 0 , 1 ] 2 . Bedregal et al. [26] defined the pointwise order of two grouping functions in a similar way. Dai et al. [33] showed that the meet operation, join operation, convex combination, and ⊛-composition of overlap and grouping functions are order preserving. But the research on the pointwise comparability of overlap and grouping functions have not been studied in details. Therefore, in this paper, we study the pointwise comparability of overlap and grouping functions involving Bustince et al. [1] and Bedregal et al. [26] generators. We present some necessary and sufficient conditions for their comparability. We also investigate order preservation of some compositions of overlap and grouping functions.
The paper is organized as follows: In Section 2, we recall the concepts of overlap/grouping functions and their order relationship. In Section 3, we study the pointwise comparability of overlap functions involving Bustince et al. [1] generators. In Section 4, we study the pointwise comparability of grouping functions involving Bedregal et al. [26] generators. In Section 5, we introduce some compositions of overlap/grouping functions and study properties preservation of these compositions. In Section 6, our researches are concluded.

2. Preliminaries

2.1. Overlap and Grouping Functions

In this section, we recall the basic theory of overlap and grouping functions. More details can be found in [1,11,26,28].
Definition 1
([1]). A bivariate function O : [ 0 , 1 ] 2 [ 0 , 1 ] is a overlap function if, for any a , b [ 0 , 1 ] , it has the following properties:
(O1)
O is commutative;
(O2)
O ( a , b ) = 0 if and only if a b = 0 ;
(O3)
O ( a , b ) = 1 if and only if a b = 1 ;
(O4)
O is non-decreasing;
(O5)
O is continuous.
Definition 2
([3]). A bivariate function G : [ 0 , 1 ] 2 [ 0 , 1 ] is a grouping function if, for any a , b [ 0 , 1 ] , it has the following properties:
(G1)
G is commutative;
(G2)
G ( a , b ) = 0 if and only if a = b = 0 ;
(G3)
G ( a , b ) = 1 if and only if a = 1 or b = 1 .
(G4)
G is non-decreasing;
(G5)
G is continuous.
Denote by O the set of all overlap functions, and G the set of all grouping functions.
Let O be an overlap function, the dual grouping function of O is defined as G O ( a , b ) = 1 O ( 1 a , 1 b ) .
Example 1
([1,26]). The following are typical examples of overlap and grouping functions, where p > 0 ,
O n m ( a , b ) = min ( a , b ) max ( a 2 , b 2 ) ;
O p ( a , b ) = a p b p ;
O m p ( a , b ) = min ( a p , b p ) ;
O M p ( a , b ) = 1 max ( ( 1 a ) p , ( 1 b ) p ) ;
O D B ( a , b ) = 2 a b a + b , i f a + b 0 , 0 , i f a + b = 0 .
G n m ( a , b ) = 1 min ( 1 a , 1 b ) max ( ( 1 a ) 2 , ( 1 b ) 2 ) ;
G p ( a , b ) = 1 ( 1 a ) p ( 1 b ) p ;
G m p ( a , b ) = 1 min ( ( 1 a ) p , ( 1 b ) p ) ;
G M p ( a , b ) = max ( a p , b p ) ;
G D B ( a , b ) = a + b 2 a b 2 a b , i f a 1 o r b 1 , 1 , i f a = b = 1 .

2.2. Orders of Overlap and Grouping Functions

Bustince et al. [1] and Bedregal et al. [26] introduced the following partial order for overlap and grouping functions, respectively.
Definition 3
([1,26]). Let f 1 , f 2 O (or both f 1 , f 2 G ),
(i)
we say that f 1 is weaker than f 2 , denote f 1 f 2 , if f 1 ( a , b ) f 2 ( a , b ) holds for all ( a , b ) [ 0 , 1 ] 2 .
(ii)
we write f 1 f 2 if f 1 f 2 and f 1 f 2 .
Proposition 1.
Let O 1 and O 2 be two overlap functions, if O 1 O 2 , then G O 2 G O 1 , where G O 1 and G O 2 are the dual grouping functions of O 1 and O 2 , respectively.
Proof. 
First O 1 O 2 means O 1 ( a , b ) O 2 ( a , b ) holds for all ( a , b ) [ 0 , 1 ] 2 . Then O 1 ( 1 a , 1 b ) O 2 ( 1 a , 1 b ) holds for all ( a , b ) [ 0 , 1 ] 2 .
Afterwards we have 1 O 1 ( 1 a , 1 b ) 1 O 2 ( 1 a , 1 b ) holds for all ( a , b ) [ 0 , 1 ] 2 . Thus G O 2 G O 1 , i.e., G O 2 ( a , b ) G O 1 ( a , b ) holds for all ( a , b ) [ 0 , 1 ] 2 . □
Example 2.
Consider the overlap and grouping functions in Example 1, we have
O n m O m p , where 0 < p 1 ;
O m p O n m , where p 3 ;
O p O m p ;
O p O D B , where p 1 ;
G m p G n m , where 0 < p 1 ;
G n m G m p , where p 3 ;
G m p G p ;
G D B G p , where p 1 .
Remark 1.
is a partial order, but not a linear order. For example, consider the O m p with p = 2 and O n m , O m p ( a , b ) = min ( a 2 , b 2 ) and O n m are incomparable since O m p ( 1 , 1 2 ) = 1 4 < O n m ( 1 , 1 2 ) = 1 2 and O m p ( 1 2 , 1 2 ) = 1 4 > O n m ( 1 2 , 1 2 ) = 1 8 .

3. Comparison of Overlap Functions

Bustince et al. [1] gave an alternative characterization of overlap functions.
Theorem 1
([1]). The bivariate function O f g : [ 0 , 1 ] 2 [ 0 , 1 ] is an overlap function if and only if
O f g ( a , b ) = f ( a , b ) f ( a , b ) + g ( a , b )
for some f , g : [ 0 , 1 ] 2 [ 0 , 1 ] satisfying the following conditions
(F1)
f and g are symmetric;
(F2)
f is non decreasing and g is non increasing;
(F3)
f ( a , b ) = 0 if and only if a b = 0 ;
(F4)
g ( a , b ) = 0 if and only if a b = 1 ;
(F5)
f and g are continuous functions.
For any overlap function O f g characterized by Equation (1), ( f , g ) is said to be the generator pair of O f g .
We give the following necessary and sufficient condition for the comparison of overlap functions characterized by different generator pairs.
Theorem 2.
Let O f 1 g 1 and O f 2 g 2 be two overlap functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then O f 1 g 1 O f 2 g 2 if and only if f 1 g 2 f 2 g 1 , i.e., for all a , b [ 0 , 1 ] ,
f 1 ( a , b ) g 2 ( a , b ) f 2 ( a , b ) g 1 ( a , b ) .
Proof. 
(⇒) From the definition of the generator pair in Equation (1), if O f 1 g 1 O f 2 g 2 , then for all a , b [ 0 , 1 ] , by Definition 3,
f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) .
From Theorem 1 (F3) and (F4), we have f 1 ( a , b ) + g 1 ( a , b ) > 0 and f 2 ( a , b ) + g 2 ( a , b ) > 0 for all a , b [ 0 , 1 ] , and we have
f 1 ( a , b ) [ f 2 ( a , b ) + g 2 ( a , b ) ] f 2 ( a , b ) [ f 1 ( a , b ) + g 1 ( a , b ) ] .
Then for all a , b [ 0 , 1 ] , it holds that
f 1 ( a , b ) g 2 ( a , b ) f 2 ( a , b ) g 1 ( a , b )
Thus f 1 g 2 f 2 g 1 .
(⇐) If f 1 g 2 f 2 g 1 , i.e., for all a , b [ 0 , 1 ] , it holds that
f 1 ( a , b ) g 2 ( a , b ) f 2 ( a , b ) g 1 ( a , b ) .
By adding f 1 ( a , b ) f 2 ( a , b ) in both sides of this inequality, we obtain
f 1 ( a , b ) f 2 ( a , b ) + f 1 ( a , b ) g 2 ( a , b ) f 1 ( a , b ) f 2 ( a , b ) + f 2 ( a , b ) g 1 ( a , b ) ,
i.e.,
f 1 ( a , b ) [ f 2 ( a , b ) + g 2 ( a , b ) ] f 2 ( a , b ) [ f 1 ( a , b ) + g 1 ( a , b ) ] .
From f 1 ( a , b ) + g 1 ( a , b ) > 0 and f 2 ( a , b ) + g 2 ( a , b ) > 0 for all a , b [ 0 , 1 ] , one has that [ f 2 ( a , b ) + g 2 ( a , b ) ] [ f 1 ( a , b ) + g 1 ( a , b ) ] > 0 for all a , b [ 0 , 1 ] .
Then by dividing both sides of the equation by [ f 2 ( a , b ) + g 2 ( a , b ) ] [ f 1 ( a , b ) + g 1 ( a , b ) ] , we get for all a , b [ 0 , 1 ]
f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) .
Thus by Definition 3, O f 1 g 1 O f 2 g 2 . □
Corollary 1.
Let O f 1 g 1 and O f 2 g 2 be two overlap functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then O f 1 g 1 O f 2 g 2 if and only if f 1 f 2 g 1 g 2 , i.e., for all ( a , b ) ( 0 , 1 ] 2 \ { ( 1 , 1 ) } ,
f 1 ( a , b ) f 2 ( a , b ) g 1 ( a , b ) g 2 ( a , b ) .
Corollary 2.
Let O f 1 g and O f 2 g be two overlap functions with generator pair ( f 1 , g ) and ( f 2 , g ) , respectively. If f 1 f 2 , i.e., f 1 ( a , b ) f 2 ( a , b ) for all a , b [ 0 , 1 ] . Then O f 1 g O f 2 g .
Corollary 3.
Let O f g 1 and O f g 2 be two overlap functions with generator pair ( f , g 1 ) and ( f , g 2 ) , respectively. If g 1 g 2 , i.e., g 1 ( a , b ) g 2 ( a , b ) for all a , b [ 0 , 1 ] . Then O f g 2 O f g 1 .
Example 3.
Consider the following functions f 1 , g 1 , f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , defined by
f 1 ( a , b ) = a b ,
f 2 ( a , b ) = a 2 b 2 ,
g 1 ( a , b ) = 1 a b ,
g 2 ( a , b ) = max ( 1 a , 1 b ) .
Obviously, they satify the conditions of Theorem 1. We also have f 2 f 1 and g 2 g 1 . Then it holds that f 2 g 2 f 1 g 1 .
From Theorem 2, we obtain O f 2 g 1 O f 1 g 2 , i.e.,
a 2 b 2 a 2 b 2 + 1 a b a b a b + max ( 1 a , 1 b )
for all a , b [ 0 , 1 ] .
Moreover,
f 1 ( a , b ) f 2 ( a , b ) = a b a 2 b 2 = 1 a 3 / 2 b 3 / 2
and
g 1 ( a , b ) g 2 ( a , b ) = 1 a b max ( 1 a , 1 b )
are incomparable since
f 1 ( 0.9 , 0.9 ) f 2 ( 0.9 , 0.9 ) = 1 0 . 9 3 1.372 < g 1 ( 0.9 , 0.9 ) g 2 ( 0.9 , 0.9 ) = 1.9
and
f 1 ( 0.1 , 0.1 ) f 2 ( 0.1 , 0.1 ) = 1 0 . 1 3 = 1000 > g 1 ( 0.1 , 0.1 ) g 2 ( 0.1 , 0.1 ) = 1.1 .
Then O f 1 g 1 ( a , b ) = a b a b + 1 a b and O f 2 g 2 ( a , b ) = a 2 b 2 a 2 b 2 + max ( 1 a , 1 b ) are incomparable because of Corollary 1.

4. Comparison of Grouping Functions

Bedregal et al. [26] gave an alternative characterization of grouping functions.
Theorem 3
([26]). The bivariate function G f g : [ 0 , 1 ] 2 [ 0 , 1 ] is a grouping function if and only if
G f g ( a , b ) = 1 f ( a , b ) f ( a , b ) + g ( a , b )
for some f , g : [ 0 , 1 ] 2 [ 0 , 1 ] satisfying the following conditions
(T1)
f and g are symmetric;
(T2)
f is non increasing and g is non decreasing;
(T3)
f ( a , b ) = 0 if and only if a = 1 or b = 1 ;
(T4)
g ( a , b ) = 0 if and only if a = b = 0 ;
(T5)
f and g are continuous functions.
For any grouping function G f g characterized by Equation (8), ( f , g ) is said to be the generator pair of G f g .
We give the following necessary and sufficient condition for the comparison of grouping functions characterized by different generator pairs.
Theorem 4.
Let G f 1 g 1 and G f 2 g 2 be two grouping functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then G f 1 g 1 G f 2 g 2 if and only if f 2 g 1 f 1 g 2 , i.e., for all a , b [ 0 , 1 ] ,
f 2 ( a , b ) g 1 ( a , b ) f 1 ( a , b ) g 2 ( a , b ) .
Proof. 
(⇒) From the definition of the generator pair in Equation (8), if G f 1 g 1 G f 2 g 2 , then for all a , b [ 0 , 1 ] , by Definition 3,
1 f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) 1 f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) .
This is
f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) .
From Theorem 3 (T3) and (T4), we have f 1 ( a , b ) + g 1 ( a , b ) > 0 and f 2 ( a , b ) + g 2 ( a , b ) > 0 for all a , b [ 0 , 1 ] , and we have
f 2 ( a , b ) [ f 1 ( a , b ) + g 1 ( a , b ) ] f 1 ( a , b ) [ f 2 ( a , b ) + g 2 ( a , b ) ] .
Then for all a , b [ 0 , 1 ] , it holds that
f 2 ( a , b ) g 1 ( a , b ) f 1 ( a , b ) g 2 ( a , b ) .
Thus f 2 g 1 f 1 g 2 .
(⇐) If f 2 g 1 f 1 g 2 , i.e., for all a , b [ 0 , 1 ] , it holds that
f 2 ( a , b ) g 1 ( a , b ) f 1 ( a , b ) g 2 ( a , b ) .
By adding f 1 ( a , b ) f 2 ( a , b ) in both sides of this inequality, we obtain
f 1 ( a , b ) f 2 ( a , b ) + f 2 ( a , b ) g 1 ( a , b ) f 1 ( a , b ) f 2 ( a , b ) + f 1 ( a , b ) g 2 ( a , b ) ,
i.e.,
f 2 ( a , b ) [ f 1 ( a , b ) + g 1 ( a , b ) ] f 1 ( a , b ) [ f 2 ( a , b ) + g 2 ( a , b ) ] .
From f 1 ( a , b ) + g 1 ( a , b ) > 0 and f 2 ( a , b ) + g 2 ( a , b ) > 0 for all a , b [ 0 , 1 ] , one has that [ f 2 ( a , b ) + g 2 ( a , b ) ] [ f 1 ( a , b ) + g 1 ( a , b ) ] > 0 for all a , b [ 0 , 1 ] .
Then by dividing both sides of the equation by [ f 2 ( a , b ) + g 2 ( a , b ) ] [ f 1 ( a , b ) + g 1 ( a , b ) ] , we get for all a , b [ 0 , 1 ]
f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) .
So we have, for all a , b [ 0 , 1 ]
1 f 1 ( a , b ) f 1 ( a , b ) + g 1 ( a , b ) 1 f 2 ( a , b ) f 2 ( a , b ) + g 2 ( a , b ) .
Thus by Definition 3, G f 1 g 1 G f 2 g 2 . □
Corollary 4.
Let G f 1 g 1 and G f 2 g 2 be two grouping functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then G f 1 g 1 G f 2 g 2 if and only if g 1 g 2 f 1 f 2 , i.e., for all ( a , b ) [ 0 , 1 ) 2 { ( 0 , 0 ) } ,
g 1 ( a , b ) g 2 ( a , b ) f 1 ( a , b ) f 2 ( a , b ) .
Corollary 5.
Let G f 1 g and G f 2 g be two grouping functions with generator pair ( f 1 , g ) and ( f 2 , g ) , respectively. If f 1 f 2 , then G f 2 g G f 1 g .
Corollary 6.
Let G f g 1 and G f g 2 be two grouping functions with generator pair ( f , g 1 ) and ( f , g 2 ) , respectively. If g 1 g 2 , then G f g 1 G f g 2 .
Example 4.
Consider the following functions f 1 , g 1 , f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , defined by
f 1 ( a , b ) = ( 1 a ) ( 1 b ) ,
f 2 ( a , b ) = ( 1 a 2 ) ( 1 b 2 ) ,
g 1 ( a , b ) = a + b 2 ,
g 2 ( a , b ) = min ( a , b ) .
Obviously, they satify the conditions of Theorem 3. We also have f 1 f 2 and g 2 g 1 . Then it holds that f 1 g 2 f 2 g 1 .
From Theorem 4, we obtain G f 2 g 2 G f 1 g 1 , i.e.,
1 ( 1 a 2 ) ( 1 b 2 ) ( 1 a 2 ) ( 1 b 2 ) + min ( a , b ) 1 ( 1 a ) ( 1 b ) ( 1 a ) ( 1 b ) + a + b 2
for all a , b [ 0 , 1 ] .
Moreover, for all ( a , b ) [ 0 , 1 ) 2 \ { ( 0 , 0 ) }
f 1 ( a , b ) f 2 ( a , b ) = ( 1 a ) ( 1 b ) ( 1 a 2 ) ( 1 b 2 ) = 1 ( 1 + a ) ( 1 + b )
and
g 2 ( a , b ) g 1 ( a , b ) = min ( a , b ) a + b 2 = 2 min ( a , b ) a + b
are incomparable since
f 1 ( 0.25 , 0.25 ) f 2 ( 0.25 , 0.25 ) = 1 1 . 25 2 = 0.64 < g 2 ( 0.25 , 0.25 ) g 1 ( 0.25 , 0.25 ) = 1
and
f 1 ( 0.1 , 0.9 ) f 2 ( 0.1 , 0.9 ) = 1 1.1 1.9 0.4785 > g 1 ( 0.1 , 0.9 ) g 2 ( 0.1 , 0.9 ) = 0.2 .
Then G f 1 g 2 ( a , b ) = 1 ( 1 a ) ( 1 b ) ( 1 a ) ( 1 b ) + min ( a , b ) and G f 2 g 1 ( a , b ) = 1 ( 1 a 2 ) ( 1 b 2 ) ( 1 a 2 ) ( 1 b 2 ) + a + b 2 are incomparable because of Corollary 4.

5. Order Preservation of Some Compositions of Overlap and Grouping Functions

In this section, we consider the following problem.
Problem 1.
Whether we have
H f 1 g 1 H f 2 g 2 , H f 3 g 3 H f 4 g 4 H ( f 1 1 f 3 ) ( g 1 1 g 3 ) H ( f 2 2 f 4 ) ( g 2 2 g 4 )
for some operations 1 and 2 of bivariate functions, where H f i g i , with i = 1 , . . . , 4 are all overlap functions or all grouping functions?
Let h 1 and h 2 be two bivariate functions, their meet, join and product operations are defined as
( h 1 h 2 ) ( a , b ) = max h 1 ( a , b ) , h 2 ( a , b ) ,
( h 1 h 2 ) ( a , b ) = min h 1 ( a , b ) , h 2 ( a , b ) ,
( h 1 × h 2 ) ( a , b ) = h 1 ( a , b ) h 2 ( a , b ) ,
for all ( a , b ) [ 0 , 1 ] 2 .
First, we prove the closures of the proposed compositions of overlap (or grouping) functions H ( f 1 f 2 ) ( g 1 g 2 ) , where { , , × } .
Lemma 1.
If f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] satisfy the following conditions (F1)–(F5) of Theorem 1. Then ( f 1 f 2 , g 1 g 2 ) , ( f 1 f 2 , g 1 g 2 ) and ( f 1 × f 2 , g 1 × g 2 ) also satisfy these conditions.
Proof. 
The cases for (F1), (F2), and (F5) are straightforward.
(F3) (⇒) If ( f 1 f 2 ) ( a , b ) = max f 1 ( a , b ) , f 2 ( a , b ) = 0 . Then, f 1 ( a , b ) = f 2 ( a , b ) = 0 , thus a b = 0 .
If ( f 1 f 2 ) ( a , b ) = min f 1 ( a , b ) , f 2 ( a , b ) = 0 . Case I, f 1 ( a , b ) = 0 then a b = 0 . Case II, f 2 ( a , b ) = 0 then a b = 0 .
If ( f 1 × f 2 ) ( a , b ) = f 1 ( a , b ) f 2 ( a , b ) = 0 . Case I, f 1 ( a , b ) = 0 then a b = 0 . Case II, f 2 ( a , b ) = 0 then a b = 0 .
(⇐) is straightforward.
(F4) (⇒) If ( g 1 g 2 ) ( a , b ) = max g 1 ( a , b ) , g 2 ( a , b ) = 0 . Then g 1 ( a , b ) = g 2 ( a , b ) = 0 , thus a b = 1 .
If ( g 1 g 2 ) ( a , b ) = min g 1 ( a , b ) , g 2 ( a , b ) = 0 . Case I, g 1 ( a , b ) = 0 then a b = 1 . Case II, g 2 ( a , b ) = 0 then a b = 1 .
If ( g 1 × g 2 ) ( a , b ) = g 1 ( a , b ) g 2 ( a , b ) = 0 . Case I, g 1 ( a , b ) = 0 then a b = 1 . Case II, g 2 ( a , b ) = 0 then a b = 1 .
(⇐) is straightforward. □
Corollary 7.
If O f 1 g 1 and O f 2 g 2 be two overlap function functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then O ( f 1 f 2 ) ( g 1 g 2 ) is an overlap function, where { , , × }
Lemma 2.
If f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] satisfy the following conditions (T1)-(T5) of Theorem 3. Then ( f 1 f 2 , g 1 g 2 ) , ( f 1 f 2 , g 1 g 2 ) and ( f 1 × f 2 , g 1 × g 2 ) also satisfy these conditions.
Proof. 
The cases for (T1), (T2), and (T5) are straightforward.
(T3) (⇒) If ( f 1 f 2 ) ( a , b ) = max f 1 ( a , b ) , f 2 ( a , b ) = 0 . Then, f 1 ( a , b ) = f 2 ( a , b ) = 0 , thus a = 1 or b = 1 .
If ( f 1 f 2 ) ( a , b ) = min f 1 ( a , b ) , f 2 ( a , b ) = 0 . Case I, f 1 ( a , b ) = 0 then a = 1 or b = 1 . Case II, f 2 ( a , b ) = 0 then a = 1 or b = 1 .
If ( f 1 × f 2 ) ( a , b ) = f 1 ( a , b ) f 2 ( a , b ) = 0 . Case I, f 1 ( a , b ) = 0 then a = 1 or b = 1 . Case II, f 2 ( a , b ) = 0 then a = 1 or b = 1 .
(⇐) is straightforward.
(T4) (⇒) If ( g 1 g 2 ) ( a , b ) = max g 1 ( a , b ) , g 2 ( a , b ) = 0 . Then g 1 ( a , b ) = g 2 ( a , b ) = 0 , thus a = b = 0 .
If ( g 1 g 2 ) ( a , b ) = min g 1 ( a , b ) , g 2 ( a , b ) = 0 . Case I, g 1 ( a , b ) = 0 then a = b = 0 . Case II, g 2 ( a , b ) = 0 then a = b = 0 .
If ( g 1 × g 2 ) ( a , b ) = g 1 ( a , b ) g 2 ( a , b ) = 0 . Case I, g 1 ( a , b ) = 0 then a = b = 0 . Case II, g 2 ( a , b ) = 0 then a = b = 0 .
(⇐) is straightforward. □
Corollary 8.
If G f 1 g 1 and G f 2 g 2 be two grouping functions with generator pair f 1 , g 1 : [ 0 , 1 ] 2 [ 0 , 1 ] and f 2 , g 2 : [ 0 , 1 ] 2 [ 0 , 1 ] , respectively. Then G ( f 1 f 2 ) ( g 1 g 2 ) is a grouping function, where { , , × }
Lemma 3.
Let a 1 , a 2 , a 3 , a 4 , b 1 , b 2 , b 3 , b 4 [ 0 , 1 ] . If a 1 b 2 a 2 b 1 and a 3 b 4 a 4 b 3 . Then
(1)
( a 1 a 3 ) ( b 2 b 4 ) ( a 2 a 4 ) ( b 1 b 3 ) .
(2)
( a 1 a 3 ) ( b 2 b 4 ) ( a 2 a 4 ) ( b 1 b 3 ) .
Theorem 5.
Let O f 1 g 1 , O f 2 g 2 , O f 3 g 3 and O f 4 g 4 be four overlap functions with generator pair ( f 1 , g 1 ) , ( f 2 , g 2 ) , ( f 3 , g 3 ) and ( f 4 , g 4 ) , respectively. If O f 1 g 1 O f 2 g 2 and O f 3 g 3 O f 4 g 4 , then
(1)
O ( f 1 × f 3 ) ( g 1 × g 3 ) O ( f 2 × f 4 ) ( g 2 × g 4 ) ;
(2)
O ( f 1 f 3 ) ( g 1 g 3 ) O ( f 2 f 4 ) ( g 2 g 4 ) .
Proof. 
From O f 1 g 1 O f 2 g 2 and O f 3 g 3 O f 4 g 4 , by Theorem 1, it hold that f 1 g 2 f 2 g 1 and f 3 g 4 f 4 g 3 . By Lemma 3(1) and (2), we have ( f 1 × f 3 ) ( g 2 × g 4 ) ( f 2 × f 4 ) ( g 1 × g 3 ) and ( f 1 f 3 ) ( g 2 g 4 ) ( f 2 f 4 ) ( g 1 g 3 ) , respectively. Thus we obtain O ( f 1 × f 3 ) ( g 1 × g 3 ) O ( f 2 × f 4 ) ( g 2 × g 4 ) and O ( f 1 f 3 ) ( g 1 g 3 ) O ( f 2 f 4 ) ( g 2 g 4 ) . □
Theorem 6.
Let G f 1 g 1 , G f 2 g 2 , G f 3 g 3 and G f 4 g 4 be four grouping functions with generator pair ( f 1 , g 1 ) , ( f 2 , g 2 ) , ( f 3 , g 3 ) and ( f 4 , g 4 ) , respectively. If G f 1 g 1 G f 2 g 2 and G f 3 g 3 G f 4 g 4 , then
(1)
G ( f 1 × f 3 ) ( g 1 × g 3 ) G ( f 2 × f 4 ) ( g 4 × g 4 ) ;
(2)
G ( f 1 f 3 ) ( g 1 g 3 ) G ( f 2 f 4 ) ( g 2 g 4 ) .
Proof. 
From G f 1 g 1 G f 2 g 2 and G f 3 g 3 G f 4 g 4 , by Theorem 4, it hold that f 2 g 1 f 1 g 2 and f 4 g 3 f 3 g 4 . By Lemma 3(1) and (2), we have ( f 2 × f 4 ) ( g 1 × g 3 ) ( f 1 × f 3 ) ( g 2 × g 4 ) and ( f 2 f 4 ) ( g 1 g 3 ) ( f 1 f 3 ) ( g 2 g 4 ) , respectively. Thus we obtain G ( f 1 × f 3 ) ( g 1 × g 3 ) G ( f 2 × f 4 ) ( g 2 × g 4 ) and G ( f 1 f 3 ) ( g 1 g 3 ) G ( f 2 f 4 ) ( g 2 g 4 ) . □

6. Conclusions

This paper studies the pointwise comparability of overlap and grouping functions, respectively. We give some necessary and sufficient conditions for the comparison of overlap functions characterized by Bustince et al. generator pairs [1] and grouping functions characterized by Bedregal et al. generator pairs [26]. We present some more general results on order preservation with respect to some compositions of overlap and grouping functions.
In this paper, we only focus on overlap and grouping functions characterized by Bustince et al. [1] and Bedregal et al. [26] generators, respectively. Naturally, a more detailed discussion of other generators of overlap and grouping functions, such as additive generators proposed by Dimuro et al. [27] and multiplicative generators proposed by Qiao and Hu [29], will be both necessary and interesting.
It was observed that there are various compositions of overlap and grouping functions. Obviously, the results presented in this paper are particular cases of order preservation with respect to some compositions of overlap and grouping functions. Therefore, it will be meaningful to further discuss order preservation of other compositions.

Funding

This research was funded by the National Science Foundation of China under Grant No. 62006168 and Zhejiang Provincial Natural Science Foundation of China under Grant No. LQ21A010001.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Bustince, H.; Fernández, J.; Mesiar, R.; Montero, J.; Orduna, R. Overlap functions. Nonlinear Anal. Theory Methods Appl. 2010, 72, 1488–1499. [Google Scholar] [CrossRef]
  2. Beliakov, G.; Pradera, A.; Calvo, T. Aggregation Functions: A Guide for Practitioners; Springer: Berlin/Heidelberg, Germany, 2007. [Google Scholar]
  3. Bustince, H.; Pagola, M.; Mesiar, R.; Hüllermeier, E.; Herrera, F. Grouping, overlaps, and generalized bientropic functions for fuzzy modeling of pairwise comparisons. IEEE Trans. Fuzzy Syst. 2012, 20, 405–415. [Google Scholar] [CrossRef]
  4. Jurio, A.; Bustince, H.; Pagola, M.; Pradera, A.; Yager, R. Some properties of overlap and grouping functions and their application to image thresholding. Fuzzy Sets Syst. 2013, 229, 69–90. [Google Scholar] [CrossRef]
  5. Elkano, M.; Galar, M.; Sanz, J.; Bustince, H. Fuzzy Rule-Based Classification Systems for multi-class problems using binary decomposition strategies: On the influence of n-dimensional overlap functions in the Fuzzy Reasoning Method. Inf. Sci. 2016, 332, 94–114. [Google Scholar] [CrossRef] [Green Version]
  6. Elkano, M.; Galar, M.; Sanz, J.; Fernández, A.; Barrenechea, E.; Herrera, F.; Bustince, H. Enhancing multi-class classification in FARC-HD fuzzy classifier: On the synergy between n-dimensional overlap functions and decomposition strategies. IEEE Trans. Fuzzy Syst. 2015, 23, 1562–1580. [Google Scholar] [CrossRef] [Green Version]
  7. Elkano, M.; Galar, M.; Sanz, J.A.; Schiavo, P.F.; Pereira, S.; Dimuro, G.P.; Borges, E.N.; Bustince, H. Consensus via penalty functions for decision making in ensembles in fuzzy rule-based classification systems. Appl. Soft Comput. 2018, 67, 728–740. [Google Scholar] [CrossRef]
  8. Santos, H.; Lima, L.; Bedregal, B.; Dimuro, G.P.; Rocha, M.; Bustince, H. Analyzing subdistributivity and superdistributivity on overlap and grouping functions. In Proceedings of the 8th International Summer School on Aggregation Operators (AGOP 2015), Katowice, Poland, 7–10 July 2015; pp. 211–216. [Google Scholar]
  9. De Miguel, L.; Gomez, D.; Rodríguez, J.T.; Montero, J.; Bustince, H.; Dimuro, G.P.; Sanz, J.A. General overlap functions. Fuzzy Sets Syst. 2019, 372, 81–96. [Google Scholar] [CrossRef]
  10. Santos, H.; Dimuro, G.P.; Asmus, T.C.; Lucca, G.; Bueno, E.; Bedregal, B.; Bustince, H. General grouping functions. In Proceedings of the 18th International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems, Lisbon, Portugal, 15–19 June 2020. [Google Scholar]
  11. Gómez, D.; Rodríguez, J.T.; Montero, J.; Bustince, H.; Barrenechea, E. N-dimensional overlap functions. Fuzzy Sets Syst. 2016, 287, 57–75. [Google Scholar] [CrossRef]
  12. Dimuro, G.P.; Bedregal, B. Archimedean overlap functions: The ordinal sum and the cancellation, idempotency and limiting properties. Fuzzy Sets Syst. 2014, 252, 39–54. [Google Scholar] [CrossRef]
  13. Asmus, T.C.; Dimuro, G.P.; Bedregal, B.; Sanz, J.A.; Pereira, S.; Bustince, H. General interval-valued overlap functions and interval-valued overlap indices. Inf. Sci. 2020, 527, 27–50. [Google Scholar] [CrossRef]
  14. Chen, Y.; Bi, L.; Hu, B.; Dai, S. General Complex-Valued Overlap Functions. J. Math. 2021, 2021, 6613730. [Google Scholar] [CrossRef]
  15. Chen, Y.; Bi, L.; Hu, B.; Dai, S. General Complex-Valued Grouping Functions. J. Math. 2021, 2021, 5793151. [Google Scholar] [CrossRef]
  16. Costa, L.M.; Bedregal, B.R.C. Quasi-homogeneous overlap functions. In Decision Making and Soft Computing; World Scientific: Singapore, 2014; pp. 294–299. [Google Scholar]
  17. Qiao, J.; Hu, B.Q. On homogeneous, quasi-homogeneous and pseudo-homogeneous overlap and grouping functions. Fuzzy Sets Syst. 2019, 357, 58–90. [Google Scholar] [CrossRef]
  18. Wang, H. Constructions of overlap functions on bounded lattices. Int. J. Approx. Reason. 2020, 125, 203–217. [Google Scholar] [CrossRef]
  19. Qiao, J. Overlap and grouping functions on complete lattices. Inf. Sci. 2021, 542, 406–424. [Google Scholar] [CrossRef]
  20. Bedregal, B.; Bustince, H.; Palmeira, E.; Dimuro, G.; Fernandez, J. Generalized interval-valued OWA operators with interval weights derived from interval-valued overlap functions. Int. J. Approx. Reason. 2017, 90, 1–16. [Google Scholar] [CrossRef] [Green Version]
  21. Dimuro, G.P.; Bedregal, B. On residual implications derived from overlap functions. Inf. Sci. 2015, 312, 78–88. [Google Scholar] [CrossRef]
  22. Dimuro, G.P.; Bedregal, B. On the laws of contraposition for residual implications derived from overlap functions. In Proceedings of the 2015 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Istanbul, Turkey, 2–5 August 2015; pp. 1–7. [Google Scholar]
  23. Dimuro, G.P.; Bedregal, B.; Santiago, R.H.N. On (G,N)-implications derived from grouping functions. Inf. Sci. 2014, 279, 1–17. [Google Scholar] [CrossRef]
  24. Qiao, J. On binary relations induced from overlap and grouping functions. Int. J. Approx. Reason. 2019, 106, 155–171. [Google Scholar] [CrossRef]
  25. Qiao, J. On (IO, O)-fuzzy rough sets based on overlap functions. Int. J. Approx. Reason. 2021, 132, 26–48. [Google Scholar] [CrossRef]
  26. Bedregal, B.; Dimuro, G.P.; Bustince, H.; Barrenechea, E. New results on overlap and grouping functions. Inf. Sci. 2013, 249, 148–170. [Google Scholar] [CrossRef]
  27. Dimuro, G.P.; Bedregal, B.; Bustince, H.; Asiáin, M.J.; Mesiar, R. On additive generators of overlap functions. Fuzzy Sets Syst. 2016, 287, 76–96. [Google Scholar] [CrossRef] [Green Version]
  28. Qiao, J.; Hu, B.Q. On interval additive generators of interval overlap functions and interval grouping functions. Fuzzy Sets Syst. 2017, 323, 19–55. [Google Scholar] [CrossRef]
  29. Qiao, J.; Hu, B.Q. On multiplicative generators of overlap and grouping functions. Fuzzy Sets Syst. 2018, 332, 1–24. [Google Scholar] [CrossRef]
  30. Klement, E.P.; Mesiar, R.; Pap, E. Triangular Norms; Kluwer: Dordrecht, The Netherlands, 2000. [Google Scholar]
  31. Klement, E.P.; Mesiar, R.; Pap, E. A characterization of the ordering of continuous t-norms. Fuzzy Sets Syst. 1997, 86, 189–195. [Google Scholar] [CrossRef]
  32. Baczyński, M.; Jayaram, B. Fuzzy Implications; Springer: Berlin/Heidelberg, Germany, 2008. [Google Scholar]
  33. Dai, S.; Du, L.; Song, H.; Xu, Y. On the Composition of Overlap and Grouping Functions. Axioms 2021, 10, 272. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Dai, S. Comparison of Overlap and Grouping Functions. Axioms 2022, 11, 420. https://doi.org/10.3390/axioms11080420

AMA Style

Dai S. Comparison of Overlap and Grouping Functions. Axioms. 2022; 11(8):420. https://doi.org/10.3390/axioms11080420

Chicago/Turabian Style

Dai, Songsong. 2022. "Comparison of Overlap and Grouping Functions" Axioms 11, no. 8: 420. https://doi.org/10.3390/axioms11080420

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop