1. The Modular Quasi-(Pseudo-)Metric Aggregation Problem
In this section we face the problem of merging a collection of modular quasi-(pseudo-)metrics as a natural extension of the problems, exposed in the preceding section, of aggregating a collection of modular (pseudo-)metrics.
To this end, let us introduce the notion of modular quasi-(pseudo-)metric aggregation function. Thus, given , a function is said to be a modular quasi-(pseudo-)metric aggregation function provided that, for each collection of modular quasi-(pseudo-)metrics defined on the same set X, the function is a modular quasi-(pseudo-)metric on X, where is given, for all , and for all , by
The next result was proved in [
18]. However, since it will be of great importance to our objective, we will include its proof for the sake of completeness.
Theorem 7. Let . If is a monotone function such that , then the following assertions are equivalent:
-
(1)
F is subadditive.
-
(2)
for all with .
-
(3)
F transforms n-triangular triplets into a 1-triangular triplet.
Proof. (1) ⇒ (2). Let
with
. Then
Notice that first inequality is derived from the monotony of F and the second one is due to the subadditivity of F.
(2) ⇒ (3). It is a straightforward verification.
(3) ⇒ (1). Consider . Then forms an n-dimensional triangular triplet, where . Since Ftransforms n-triangular triplets into a 1-triangular triplet we deduce that is a 1-triangular triplet. So and, thus, F is subadditve. □
Next we focus our attention on getting a characterization of modular quasi-(pseudo-)metric aggregation functions. With this aim, we note that every modular quasi-(pseudo-)metric aggregation function is a modular (pseudo-)metric aggregation function such as the following result shows.
Proposition 1. Let . If is a modular quasi-(pseudo-)metric aggregation function, then F is a modular (pseudo-)metric aggregation function.
Proof. Consider a collection of modular (pseudo-)metrics
on a non-empty set
X. Then
is a collection of modular quasi-(pseudo-)metrics on
X and, thus,
is a modular quasi-(pseudo-)metric on
X. Moreover, for all
and for all
, we have that
since
for all
. So
is a modular (pseudo-)metric on
X. Whence we conclude that
F is a modular (pseudo-)metric aggregation function. □
In the ligt of the preceding result, Theorems 5, 6 and 7 we immediately obtain the following statements.
Proposition 2. Let and let be a modular quasi-(pseudo-)metric aggregation function. Then the following assertions hold:
-
(1)
, F is monotone and subadditive.
-
(2)
and for all with .
-
(3)
, F is monotone and transforms n-triangular triplets into 1-triangular triplet.
The characterization of modular quasi-pseudo-metric aggregation functions can be sated as follows:
Theorem 8. Let and let be a function. Then the following assertions are equivalent:
-
(1)
F is a modular quasi-pseudo-metric aggregation function.
-
(2)
F is a modular pseudo-metric aggregation function.
-
(3)
, F is monotone and subadditive.
-
(4)
and for all with .
-
(5)
, F is monotone and transforms n-triangular triplets into 1-triangular triplets.
Proof. (1) ⇒ (2). follows from Proposition 1. The equivalences (2) ⇔ (3) ⇔ (4) ⇔ (5) are provided by Theorem 5. Now we prove that (4) ⇒ (1). To this end, consider a collection
of modular quasi-pseudo-metrics on a non-empty set
X. Then
for all
, for all
and for all
. Whence we obtain that
satisfies that
with
,
and
for all
. It follows that
Hence, condition (MQPM2) is satisfied. We still need to verify that condition (MQPM1) is also satisfied. Since
for all
and for all
we have that
Therefore is a modular quasi-(pseudo-)metric on X and, hence, F is a modular quasi-(pseudo-)metric aggregation function. □
The fact that every modular metric aggregation function is always a modular pseudo-metric aggregation function provides the following consequence.
Corollary 1. Let and let be a function. If F is a modular metric aggregation function, then F is a modular quasi-pseudo-metric aggregation function.
Proof. It follows immediately from Theorems 6 and 8. □
It seems natural to wonder whether the converse of the preceding corollary is also true. Nevertheless, the answer to the posed question in negative. Indeed, observe that the classes of modular quasi-pseudo-metric aggregation functions and modular pseudo-metric aggregation functions are the same. Then the function constantly equals 0, as commented in Section 0, satisfies all assumptions in the statement of Theorem 8 and, thus, it is a modular quasi-pseudo-metric aggregation function. However, when which implies that it does not satisfy all conditions in the statement of Theorem 6 and, hence, it is not a modular metric aggregation function.
Next we provide several examples of modular quasi-pseudo-metric aggregation functions.
Example 1. Let . The following functions are, for all , modular quasi-pseudo-metric aggregation functions:
-
(1)
-
(2)
.
-
(3)
.
-
(4)
for all .
-
(5)
with .
The following examples show functions which are not modular quasi-pseudo-metric aggregation functions.
Example 2.
Let . Define the functions as follows:
It is evident that F is not monotone, since with . Theorem 8 warranties that F is not a modular quasi-pseudo-metric aggregation function.
Example 3.
Let the function defided as follows:
It is clear that , so Theorem 8 ensures that F is not a modular quasi-pseudo-metric aggregation function.
Once the modular quasi-pseudo-metric aggregation problem has been studied, in the following we face a refinement of the aforementioned problem. Concretely, we try to describe those functions that are able to merge a collection of modular quasi-metrics into a single one. Accordingly, we are interested in getting an appropriate version of Theorem 8 extending Theorem 6.
The result below will play a crucial role in order to achieve our target. It must be stressed that it is an slight adaptation of a result given in [
9]. However, we have decided to include the proof, which remains the same, in order to help the reader.
Lemma 1. Let and let be a subadditive function. Then the following assertions are equivalent:
-
(1)
There exists satisfying the following: for each with we have that .
-
(2)
If such that , then .
Proof. (1) ⇒ (2). It is obvious.
(2) ⇒ (1). Suppose for the purpose of contradiction that for each there exists with and . From the fact that F is subadditive we deduce that . Whence we obtain the existence of such that and, in addition, . Nevertheless, for all , which is a contradiction because . □
Taking into account that Proposition 1 gives that every modular metric aggregation function is in fact a modular quasi-metric aggregation function we have the following characterization.
Theorem 9. Let and let be a function. The following assertions are equivalent:
-
(1)
F is a modular quasi-metric aggregation function.
-
(2)
F is a modular metric aggregation function.
-
(3)
, F is monotone and subadditive. Moreover, if and , then for some
-
(4)
and, in addition, for all with . Moreover, if and , then for some
-
(5)
, F is monotone and transforms n-triangular triplets into a 1-triangular triplet. Moreover, if and , then for some
Proof. (1) ⇒ (2). follows from Proposition 1. The equivalences (2) ⇒ (3) ⇔ (4) ⇔ (5) are provided by Theorem 6. Now we prove that (4) ⇒ (1). To this end, consider a collection of modular quasi-metrics on a non-empty set X. The same arguments to those applied to the prove of Theorem 8 gives that satisfies condition (MPQM2). It remains to prove that condition (MQM1) is hold.
Since
for all
and for all
we have that
Now assume that we have that for any and for all . Then and . By Lemma 1, there exists such that for all . The fact that is a modular quasi-metric on X yields that . So F is a modular quasi-metric aggregation function. □
The following example gives instances of modular quasi-metric aggregation functions.
Example 4. Let . The following functions are, for all , modular quasi-metric aggregation functions:
-
(1)
. Observe that this instance contains the class of weighted arithmetic means, and thus the arithmetic mean (see [19]).
-
(2)
.
-
(3)
for all . This instance contains those root-mean-powers such that (see [19]).
-
(4)
with for , where is the ith largest of the . Of course OWA operators with decreasing weights belong to this class of functions (see, for instance, [19,20]).
-
(5)
with .
-
(6)
with .
Example 2 again shows a function which is not a modular quasi-metric aggregation function, since it is not monotone. In the same way, the fuction exposed in Example 3 is not a modular quasi-metric aggregation function. Notice that in the aforementioned example the image of is not zero.
Inspired by Example 2 we give a method to construct modular quasi-(pseudo-)metric aggregation functions in the following result.
Proposition 3.
Let and be a monotone and subadditive function. Consider the function defined by
Then the following assertions hold:
-
(1)
is a modular quasi-pseudo-metric aggregation function provided that .
-
(2)
is a modular quasi-metric aggregation function provided that and that F satisfies the following property: if and , then for some
Proof. We first show that G is monotone. Let such as . Indeed, let us distinguish two possible cases.
- Case 1.
There exists such that . Then and, thus, . So .
- Case 2.
for all . Then .
Next we prove that G is subadditibve. With this aim, consider . Again, we distinguish two possible cases:
- Case 2.
There exists such that either or . Then , and either or . So .
- Case 2.
and for all . Then and .
Therefore G is subadditve.
Assume that . Then and, thus, Theorrem 8 gives that G is a modular quasi-pseudo-metric aggregation function. Finally, assume that F satisfies the property: if and , then for some . Suppose that . It follows that it must necessarily that . Then for some . Consequently, by Theorem 9, we obtain that F is a modular quasi-metric aggregation function. □
The fact that a function satisfying all assumptions in the statement of Proposition 3 is either a quasi-pseudo-metric aggregation function or a quasi-metric aggregation function, suggests us to explore the relationship between the aforementioned functions and the modular quasi-(pseudo-)metric aggregation functions.
First of all we show that there are quasi-(pseudo-)metric aggregation functions which are not modular quasi-(pseudo-)metric aggregation functions. The following example warranties such a statement.
Example 5.
Let be the function defined by
Clearly F satisfies all assumptions in Theorem 3 and thus, in Theorem 4. Whence we deduce that F is a quasi-(pseudo-)metric aggregation function. Now consider the collection of modular quasi-(pseudo-)metrics on where for all with for all and for all and for all such that . Then is not a modular (quasi-)pseudo-metric aggregation function because is not defined (observe that the value is not defined).
Notice that Example 5 also shows that there are (pseudo-)metric aggregation functions which are not modular (pseudo-)metric aggregation functions. This fact was not studied in [
18].
The example below gives an instance of modular quasi-(pseudo-)metric aggregation functions which is not a quasi-(pseudo-)metric aggregation function.
Example 6. Let . Consider the function defined by and for all . It is a simple matter to check that F is a modular quasi-(pseudo-)metric aggregation function but is not a quasi-(pseudo-)metric aggregation function.
Notice that Example 6 also shows that there are modular (pseudo-)metric aggregation functions which are not a (pseudo-)metric aggregation function. This fact was not explored in [
18].
The instances of modular quasi-metric aggregation function given in Example 4 inspires the following method to construct such functions.
Proposition 4. Let be a subadditive, monotone function such that if and only if . Let be a function such that and satisfying the following conditions:
-
(1)
If , then .
-
(2)
whenever .
If the function is subadditive, then the function defined by for all is a modular quasi-metric aggregation function.
Proof. The subadditivity of gives that subadditivity of F. Moreover, the monotony of F is directly derived from the monotony of g and condition (2). Furthermore, . Now, assume that there is such that . Then . Hence, . It follows, from condition (1), that . By Theorem 9 we have that F is a modular quasi-metric aggregation function. □
The next example shows that the “separation” condition on g can not be deleted from the statement of Proposition 4.
Example 7. Consider the function given by . Then W satisfies all assumptions in the statement of Proposition 4. Fix . Define the function by for all . The g is subadditive, monotone and satisfies that . However, but , where stands for the element of with the th coordinate as 0 and the jth coordinate with as 1. Clearly, the function given by for all fulfills that and, as a consequence, it is not a modular quasi-metric aggregation function.
The next result clarifies when a modular (quasi-)pseudo-metric aggregation function is also a (quasi-)pseudo-metric aggregation function. In order to state it we will make use of the notion of finite modular quasi-pseudo-metric aggregation function. We will say that a modular (quasi-)(pseudo-)metric aggregation function is a finite modular (quasi-)(pseudo-)metric aggregation function provided that, for each collection of modular (quasi-)(pseudo-)metrics defined on the same set X such that for all , for all and for all , the function is a modular (quasi-)(pseudo-)metric on X with for all and for all .
Theorem 10. Let and let be a modular quasi-pseudo-metric aggregation function. The following assertions are equivalent.
-
(1)
F is a finite modular quasi-pseudo-metric aggregation function.
-
(2)
F is a finite modular pseudo-metric aggregation function.
-
(3)
is a quasi-pseudo-metric aggregation function.
-
(4)
is a pseudo-metric aggregation function.
-
(5)
for some for some .
Proof. (1) ⇔ (2). It is evident.
(1) ⇒ (3). Consider a collection of quasi-pseudo-metrics defined on a non-empty set X. Define on X the collection by for all and for all . Then is a collection of modular quasi-pseudo-metrics on X. Since F is a modular quasi-pseudo-metric aggregation function we have that is a modular quasi-pseudo-metric on X. Moreover, we have, on the one hand, that for all and for all and, on the other hand, that for all and for all . Then is a quasi-pseudo-metric on X with for all . Whence we deduce that is a quasi-pseudo-metric on X. Therefore is a quasi-pseudo-metric aggregation function.
(3) ⇔ (4). The eqiuvalence is guaranteed by the fact that is monotone and by Theorems 4 and 1.
(4) ⇒ (5). For the purpose of contradiction, assume that there exists such that and for all . Define the collection of pseudo-metrics on a non-empty set X by for all , where is the discrete pseudo-metric on X. Since is a pseudo-metric aggregation function. It follows that is a pseudo-metric on X. Hence for all . However, let with , which is a contradiction.
(5) ⇒ (1). It is immediate, since . □
Similar arguments apply to the quasi-metic case.
Theorem 11. Let and let be a modular quasi-metric aggregation function. The following assertions are equivalent.
-
(1)
F is a finite modular quasi-metric aggregation function.
-
(2)
F is a finite modular metric aggregation function.
-
(3)
is a quasi-metric aggregation function.
-
(4)
is a metric aggregation function.
-
(5)
for some for some .
In the light of the preceding result, it is clear that every finite modular quasi-(pseudo-)metric aggregation function merge a collection of modular quasi-(pseudo-)metrics which do not take the value into a modular quasi-(pseudo-)metric that also does not take the value. This is the reason for the name. The functions (2), (3), (4) and (5) given in Example 1 are instances of finite modular quasi-pseudo-metric aggregation functions. Nevertheless, the function (1) provided in the aforementioned example is a modular quasi-pseudo-metric aggregation function that is not finite.
It must be pointed out that Theorems 10 and 11 stated in the modular framework are surprising due to the fact that there are (pseudo-)metric aggregation functions which are not quasi-(pseudo-)metric aggregation functions as exposed in Section 0.
It seems interesting to stress that the function (5) in Example 1 and (5) and (6) in Example 4 are instances of modular (quasi-)(pseudo-)metric aggregation functions which always merge a collection of modular (quasi-)(pseudo-)metrics into a modular (quasi-)(pseudo-)metric that does not take the value. This fact inspires the possibility of describing such kind of functions.
In the following we characterize such functions. Before stating the characterization, let us recall that, given , will denote the element of with the ith coordinate as a and the jth coordinate with as 0.
Proposition 5. Let and let be a modular (quasi-)(pseudo-)metric aggregation function. Then the following assertions are equivalent.
-
(1)
If is a collection of modular (quasi-)(pseudo-)metrics defined on a non-empty set X, then for all and for all .
-
(2)
.
-
(3)
for all .
Proof. (1) ⇒ (2). For the purpose of contradiction we assume that . Now consider a non-empty set X (with at least two different elements) and the collection of modular (quasi-)(pseudo-)metrics on X such that for all where w is defined, for all by if and if . Then is a modular (quasi-)(pseudo-)metric such that provided that . Whence we have that , which is a contradiction. So .
(2) ⇒ (3). Since F is a modular (quasi-)(pseudo-)metric aggregation function we have that it is monotone. Thus for all .
(3) ⇒ (2). Since F is a modular (quasi-)(pseudo-)metric aggregation function we have that it is subadditive. Hence .
(2) ⇒ (1). Let be a collection of modular (quasi-)(pseudo-)metrics defined on a non-empty set X. Then is a modular (quasi-)(pseudo)-metric on X. Since F is monotone and the fact that for all , for all and for all , we obtain that . □
We end this section exploring a question that arise in a natural way. Since every modular (quasi-)(pseudo-)metric aggregation function fuses a collection of modular (quasi-)(pseudo-)metrics into a single one, it seems natural to wonder the following question: Do this type of functions preserve modular (quasi-)(pseudo-)metrics? Notice that by preserving we mean that when all modular (quasi-)(pseudo-)metrics in the collection to be fused are the same then the aggregation function gives such a modular (quasi-)(pseudo-)metric as the aggregated one.
The concept below plays a central role in order to answer the posed question.
Given
, a function
has
as an idempotent element if
([
19]). In addition,
F is said to be idempotent if it has every element in
as an idempotent element, that is
for all
.
In the light of the preceding notion, the result below answers the query.
Theorem 12. Let and let X be a non-empty set. If is modular (quasi-)(pseudo-)metric aggregation function, then the following assertions are equivalent:
-
(1)
for all modular (quasi-)(pseudo-)metric on X.
-
(2)
F is idempotent.
Proof. (1) ⇒ (2). Let
. Fix
. Consider the modular (quasi-)(pseudo-)metric on a non-empty set
X given by
Then i a modular (quasi-)(pseudo-)metric on X and . So, taking a , we obtain that . Whence we conclude that F is dempotent.
(2) ⇒ (1). Consider modular (quasi-)(pseudo-)metric w on a non-empty set X. Since F is idempotent we have that for all and for all . So as claimed. □