2. Join Spaces and Connections with Lattices
In this section, we present the join space notion and we analyze some connections with lattice theory.
Let A function is called a hyperoperation, where denotes the set of nonempty subsets of H.
If are subsets of H, then
The structure
is a
hypergroup if for all
of
H we have
For all elements , denote
Definition 1. A hypergroup is called a join space if it is commutative and for all of H, we have In other words,
The condition (1) is often called the join space condition.
Join spaces were defined by W. Prenowitz. He and J. Jantosciak applied them in both Euclidean and non-Euclidean geometry, see [
16]. Using join spaces, descriptive, projective, and spherical geometry were subsequently rebuilt.
Join spaces can be also studied in connections with binary relations, fuzzy sets, or rough sets, see [
2,
17].
We present here some examples of join spaces:
Example 1. Let H be a non-empty set. If R is an equivalence relation on it, then denote the equivalence class of by and define the next hyperoperation on H: Then, is a join space.
For all we have whence is a commutative hypergroup. Moreover, if and there is such that then
If then Similarly, if then
If and then and whence hence Therefore, is a join space.
Example 2. Let be a commutative group. For all consider a nonempty set such that if , then
Set and
For all we define Then, is a join space.
Indeed, it can be checked that and
Now, if are such that then and , whence and .
Hence, , so Thus, is a join space.
Example 3. Consider a hypergroup and a commutative group. Consider a family of nonempty sets, such that and for We define the next hyperoperation on Then, is a join space if and only if is a join space.
Let be a join space.
We have , since is a group. The associativity law holds.
Moreover, if then
Similarly, if then Thus, , whence .
If , then
If , then we use the fact is a join space.
Therefore, is a join space.
Conversely, suppose that is a join space. If and , then , whence which means that Thus, is a join space.
The study of algebraic hypergroups and connections with lattices and ordered sets was initiated by J. Mittas and then by M. Konstantinidou and K. Serafimidis, Ch. Massouros, G. Massouros, and later by Ath. Kehagias. Connections between ordered sets, quasi-orders, and hypergroups were also studied by Chvalina.
In what follows, we present some connections with lattice theory, see [
18]. Two important classes of lattices are characterized using hypergroups: distributive and modular lattices, see [
9,
10,
19].
Connection 1. In [
9],
J.Varlet provided the following characterization of distributive lattices:
Consider the next hyperoperation on a lattice :
Theorem 1. is a distributive lattice if and only if is a join space.
In [
19], we considered and analyzed a family of hyperoperations
defined as follows.
Let
be arbitrary. For all
set
Theorem 2. If the lattice is distributive and , then is a join space.
We mention some important steps from the proof of this theorem:
In order to prove “⊇”, we consider an arbitrary element and we set Thus, whence
Now, we consider that satisfy and
Set We obtain .
Hence, and so is a join space.
Connection 2. The next example of a join space is useful to characterize modular lattices.
Let be a lattice. In [8] H. Nakano introduced the following hyperoperation on L: Later, S. Comer [
10] showed that:
Theorem 3. is a modular lattice if and only if is a join space.
Another interesting proof of the above theorem is given in [
2]. Several properties of this join space were presented in [
13].
In [
19], a new family of hyperoperations determined by a lattice is analyzed. For all
set
Notice that , since
For set and denote by the restriction of to .
Theorem 4. Let . If is modular, then is a join space.
Similar results can be obtained by considering the hyperoperation:
The hyperproduct is not empty since
Connection 3. Another connection between join spaces and lattices was highlighted by Tofan and Volf [14], as follows: If is a lattice and is a function, such that is a distributive sublattice of , then define We obtain a commutative hypergroup .
Indeed, in the above conditions, the next equality is checked:
Moreover, we shall prove here the next result, as follows:
Theorem 5. The next statements are equivalent:
Proof. First, let us check that the join space condition is satisfied for a distributive sublattice Let : ,
Then, we shall check that there is
Since and
, according to the distributivity, it follows that
From here we obtain
Similarly, we have
Hence,
Therefore,
Now, let us note that the reciprocal statement also holds: if the join space condition is satisfied, then is a distributive sublattice.
Indeed, if
is not distributive, then it will contain a sublattice
and
of
are not comparable two by two.
In both situations,
, since
and
Thus, and a contradiction. Thus, the sublattice is distributive. □
Canonical hypergroups are an important class of join spaces and were introduced by J. Mittas [
20]. They are the additive structures of Krasner hyperrings and were used by R. Roth to obtain results in the finite group character theory, see [
21]. McMullen and Price studied finite abelian hypergroups over splitting fields [
22].
More recent studies of canonical hypergroups were conducted by C and G. Massouros (in connection with automata), P. Corsini (sd-hypergroups), and K. Serafimidis, M. Konstantinidou, and J. Mittas, while feebly canonical hypergroups were analyzed by P. Corsini and M. De Salvo.
Canonical hypergroups are exactly join spaces with a scalar identity
e, which means that
. Obviously, commutative groups are canonical hypergroups. Other examples of canonical hypergroups are given in [
3].
More general structures were also considered, namely polygroups, also called quasi-canonical hypergroups, by Bonansinga, Corsini, and Ch. Massouros. Comer analyzed the applications of polygroups in the theory of graphs, relations, Boolean, and cylindrical algebras.
A particular type of polygroup, namely the hypergroup of bilateral classes, was investigated by Drbohlav, Harrison. and Comer. Polygroups satisfy the same conditions as canonical hypergroups, with the exception of commutativity.
In the next two sections, we associate join spaces with chains and we analyze when they are isomorphic. Moreover, a combinatorial problem is presented: we calculate how many isomorphism classes of join spaces are.
3. Join Spaces Associated with a Chain: The Finite Case
In what follows, we associate a join space structure with a chain, through a function. We then study under what conditions such join spaces, considered for different functions, are isomorphic, for the finite case.
Let
H be finite and
where
C is a chain. Consider the next hyperoperation on
H:
We have
According to Theorem 5 or by a direct check, we then utilize the following theorem.
Theorem 6. The structure is a join space.
Set
We define the next equivalence relation on
H:
Denote and order as follows: for
We denote where
For all set . We have
Denote the ordered partition of n into s parts and where
Theorem 7. If are two maps, then or .
Proof. “⇐”
Suppose Set where for all ,
Set ,
We order H as follows:
For
we have
Consider the map:
We have
which means that
Suppose now that
We have where
Moreover, , and with
Consider the function: We obtain
.
Therefore,
“⇒”
Let be defined as follows: such that Similarly, is defined, where
Denote the isomorphism by .
Denote the set
by
. For all
we have
On the other hand,
For every and every we have
Consider the function
We obtain
is injective:
Indeed, if then . Hence, for every there exists such that whence which means that is injective.
is surjective:
Indeed, for each , there is for which since
Therefore, is a bijective function from to
Particularly,
We have
So,
whence it follows that
We have
Hence,
Moreover, for all we have
From
it follows that
so
Denote by B the set of bijections of defined to itself.
We show that or Denote by
For , we have
If and we suppose that , then so which is a contradiction.
Similarly, for we obtain a contradiction.
Suppose now that there is
such that
We obtain
whence
Therefore,
which means that
If
then
whence
Thus,
that is
Hence, or □
Now we calculate how many isomorphism classes for join spaces can be constructed in this way.
Denote by the quotient set which contains classes of join spaces associated with maps .
Theorem 8. (i) If then
- (ii)
If then .
Proof. Denote by the set of ordered partitions of n.
According to [
23], we have
Let us number the symmetrical ordered partitions of n, that is partitions for which
Set the set of all symmetrical ordered partitions of n.
We have the two cases:
Case 1.
If and if , then or , where and
According to [
23], for all
t we have
Hence,
Case 2.
If and if , then or , where and
We obtain
According to [
23], for all
t we have
Therefore, we can conclude:
□
4. Join Spaces Associated with a Chain: The General Case
In this section, we consider an arbitrary set H and we analyze when the join spaces associated with a chain are isomorphic.
Let us present first the context:
Let
and consider the equivalence relation on
H:
We order as follows: for
Denote and by for all
Since C is a chain, it follows that is a chain, too.
Moreover, for all denote
If
, then
Similarly,
and for all
denote
We have that
is a chain, too.
Theorem 9. If are two functions, then if and only if there exists a strictly monotonous bijection
Proof. “⇐”
For all Define
as follows:
where we choose
. Now,
,
we have
If is strictly increasing, then
Since
is a bijective function, we have
Hence,
If is strictly decreasing, then
Since
is a bijection, we have
Hence,
Therefore, is an isomorphism.
“⇒”
Denote by the isomorphism from to Set and , where for all , we obtain and
For
we obtain
, whence
Hence,
Define by : where
We check that is a bijective function.
Suppose that there are such that
Thus, , which is a contradiction with
On the other hand, since
we obtain
whence
Hence,
is bijective and
Thus,
Hence, Let us prove now that is strictly monotonous.
If are elements of I, then we denote and for then
If
and
, then
whence
For we have or
If , then is strictly increasing on Indeed, it follows from (3).
Similarly, if , then is strictly decreasing on
Therefore, we obtain the thesis. □
Let us present some examples.
Example 4. If , where is a real interval and , then , whence and φ is the identity function. Hence,
Example 5. If , where is the real number set and , then , and again , φ is the identity function and
Example 6. If are such that , , such that , , and so on.
In general, for all .
Then, and is a strictly decreasing function. Hence,