hoskova.dvi @ Applied General Topology c© Universidad Politécnica de Valencia Volume 6, No. 1, 2005 pp. 57-65 Abelization of join spaces of affine transformations of ordered field with proximity Šárka Hošková Abstract. Using groups of affine transformations of linearly ordered fields a certain construction of non-commutative join hypergroups is presented based on the criterion of reproducibility of semi-hypergroups which are determined by ordered semigroups. The aim of this paper is to construct the abelization of the non-commutative join space of affine transformations of ordered fields. A construction of commutative weakly associative hypergroup (Hν-group) is made and a proximity is defined on this structure. 2000 AMS Classification: 20F60, 20N20 Keywords: Transformation group, join space, abelization, hyperoperation, hyperstructures, weak associativity Transformation groups which represent the classical and developing disci- pline are situated in the intersection of several parts of mathematical struc- tures. Transformation groups in the discrete approach create in a very natural way commutative hyperstuctures, which are in addition join spaces. On the other hand, groups of affine transformations of the field create naturally a transformation group on the supporting set of this field. More precisely: Let (F, +, ·) be a field, A(F) be a group of affine transformations of F of the form ϕa,b(x) = ax + b; x ∈ F , where the coefficient a ∈ F is different from zero. Then we construct a discrete transformation group (X, T, π) by putting X = F ; T = A(F) and the action π we define by π(x; ϕa,b) = ϕa,b(x). Clearly the identity axiom and homomorphism axiom are satisfied. In this paper hyperstructures on groups of affine transformations of ordered fields are constructed. These affine transformations are represented by ordered pairs of elements of a given field. During the construction of final hyperstruc- tures upper ends of the products of pairs of elements of a given field are used. This idea is adopted from the functorial assignment of a commutative hy- pergroup to an arbitrary transformation (discrete) group. We will describe this 58 Š. Hošková construction in more details. Let G = (X, T, π) be a transformation group, (i.e. X-phase set, T -phase group, π- action (projection): X × T → X). For any pair x, y ∈ X we define x ∗G y = π(x, T ) ∪ π(y, T ) = {π(x, t); t ∈ T } ∪ {π(y, t); t ∈ T }. It is easy to show that (X, ∗G) is an extensive commutative hypergroup— moreover a join space (see the definition below). In the paper [1] the authors studied the non-commutative join hypergroups of affine transformations of ordered fields. The aim of this paper is to abelize this structure using the construction described in papers [7, 14]. Recall that an ordering on a field (F, +, . ) (with the zero element 0 F and the unit 1 F ) is given by the choice of a set P ⊆ F (called the positive cone of the ordering) which satisfies the following axioms: (1) P + P ⊆ P , (2) P . P ⊆ P , (3) P ∪ (−P) = F , (4) P ∩ (−P) = {0 F } . As usual, we define a binary relation ≦P on the set F by x ≦P y (resp. x < P y) if y − x ∈ P(resp. y − x ∈ Ṗ , where Ṗ = P r {0 F }). The relation ≦P results in a total ordering of the elements of F . In detail, the relation ≦P on F is reflexive, anti symmetrical and transitive, compatible in the usual sense with commutative binary operations addition “+” and multiplication “ . ”, and satisfies the law of trichotomy (for any pair x, y ∈ F exactly one of the following possibilities occurs: x = y or x < P y or y < P x). Note that an ordered field (F, +, . , ≦P ) has necessarily the characteristic equal to zero. As usual, P ∗(S) denotes the system of all nonempty subsets of S. Let (F, +, . ) be a field of the characteristic zero. An affine transformation f : F → F of the form f(x) = a . x + b, a, b ∈ F is uniquely represented by an ordered pair of its coefficients [a, b] ∈ F × F . We shall consider non-constant transformations only, i.e., transformations f(x) = a.x + b satisfying the condi- tion a 6= 0 F . So, let us denote A(F) = (F r {0 F }) × F and define a binary operation “·” on the set A(F) by the rule [a, b] · [c, d] = [a.c, a.d + b] , which corresponds to the usual composition of affine transformations of F . It is easy to see that ( A(F), · ) is a non-commutative group with the identity [1 F , 0 F ] and inverse elements of the form [a, b]−1 = [a−1, −a−1.b]. Denote by K the subset {[a, b]; a, b ∈ F, a > P 0 F } of A(F) = (F r{0 F })×F . It is easy to see that (K, ·) is a subgroup of the group ( A(F), · ) . Define a binary relation “≦” on the set K by the rule: [a, b] ≦ [c, d] for [a, b], [c, d] ∈ K whenever a = c and b ≦P d. Then evidently “≦” is an ordering on K and we have Proposition 1 ([1]). If K = {[a, b]; a, b ∈ F, a >P 0F }, then (K, ·) is an ordered group such that it is a normal subgroup of the group ( A(F), · ) . Abelization of join spaces 59 Recall that a hypergroupoid is a pair (H, ·) where H is a (nonempty) set and · : H × H → P∗(H) is a binary hyperoperation on the set H. If a · (b · c) = (a · b) · c for all a, b, c ∈ H , (associativity) , then (H, ·) is called a semihypergroup. A semihypergroup (H, ·) is said to be a hypergroup if the following axiom a · H = H = H · a for all a ∈ H , (reproduction axiom) , is satisfied—see e.g. [3],[4]. Here, for A, B ⊆ H, A 6= ∅ 6= B we define A · B = ⋃ {a.b; a ∈ A, b ∈ B}. Moreover, for subset A and B of H, it becomes convenient to use the relational notation A ≈ B (read A meets B) to assert that A and B have an element in common, that is, that A ∩ B 6= ∅ ([16], p. 79). A hypergroup (H, ·) is called a transposition hypergroup if it satisfies the transposition axiom: For all a, b, c, d ∈ H the relation b\a ≈ c/d implies a · d ≈ b · c , where b\a = {x ∈ H; a ∈ b · x}, c/d = {x ∈ H; c ∈ x · d}. A commutative transposition hypergroup (H, ∗) is called a join space ([15]). The hyperoperation ⋆: H × H → P∗(H) is called weakly associative hyper- operation if ( a ⋆ (b ⋆ c) ) ∩ ( (a ⋆ b) ⋆ c ) 6= ∅ for any triad a, b, c ∈ H. A weak semihypergroup (Hν-semigroup) is a set H (H 6= ∅) equipped with a weakly associative hyperoperation. A Hν-semigroup is called a weak hypergroup (Hν-group) if moreover the reproduction axiom is satisfied for any a ∈ H ([18]). Proposition 2 ([5] Theorem 1). Let (S, ·, ≦) be an ordered semigroup. A binary operation ∗ : S × S → P∗(S) defined by x ∗ y = [x . y)≦ (= {t ∈ S; x . y ≦ t}) for any pair x, y ∈ S is associative. Then we have 1◦ The semi-hypergroup (S, ∗) is commutative if and only if the semigroup (S, ·) is commutative. 2◦ For the ordered semigroup (S, ·, ≦) the following conditions are equivalent: (i) for any pair of elements x, y ∈ S there exists a pair z, z′ ∈ S such that y · z ≦ x, z′ · y ≦ x, (ii) the semihypergroup (S, ∗) satisfies the reproduction condition (i.e. t∗S = S = S ∗ t for any t ∈ S), hence it is a hypergroup. Lemma 1 ([7, 14]). Let (H, ·) be a hypergroupoid. Define a hyperoperation “⋆” on the diagonal ∆ H as follows: [x, x] ⋆ [y, y] = D(x · y ∪ y · x) = { [u, u]; u ∈ x · y ∪ y · x } for any pair [x, x], [y, y] ∈ ∆ H . Then the following assertions hold: 1o For any hypergroupoid (H, ·) we have that (∆ H , ⋆) is a commutative hyper- groupoid. 60 Š. Hošková 2o If (H, ·) is a weakly associative hypergroupoid, then the hypergroupoid (∆ H , ⋆) is weakly associative, as well. 3o If (H, ·) is a quasi-hypergroup, the hypergroupoid (∆ H , ⋆) also satisfies the reproduction law, i.e., it is a quasi-hypergroup. Define a hyperoperation ∗: K × K → P∗(K) by [a, b] ∗ [c, d] = { [x, y]; [a, b] · [c, d] ≦ [x, y] } = { [a.c, y]; a.d + b ≦P y } . From Proposition 2 it is evident, that the hypergroupoid (K, ∗) is non-commutative hypergroup. The hypergroup (K, ∗) will be called to be determined by the or- dered group (K, ·, ≦). The hypergroup (K, ∗) is non-commutative, therefore we will abelize it. Let us define the set ∆ K = { [ [a, b], [a, b] ] ; [a, b] ∈ K } and a hyperoperation ⋆ : ∆ K × ∆ K → P∗(∆ K ) by [ [a, b], [a, b] ] ⋆ [ [c, d], [c, d] ] = (0.1) = { [ [x, y], [x, y] ] ; [x, y] ∈ ([a, b] ∗ [c, d]) ∪ ([c, d] ∗ [a, b]) } = { [ [x, y], [x, y] ] ; [x, y] ∈ { [a.c, u]; a.d + b ≦P u } ∪ { [c.a, v]; c.b + d ≦P v } } = { [ [x, y], [x, y] ] ; x = a.c ∧ z ≦ y) } , where z = min{a.d + b; c.b + d}. Theorem 1. The hyperstructure (∆ K , ⋆) is commutative weakly associative hypergroup, simply Hν-group. Proof. Recall that the field F is commutative, so the multiplication is commu- tative, thus a.c = c.a. It is evident, that (∆ K , ⋆) ∼= (M, ⋄), where M = { [a, b, a, b]; [a, b] ∈ K } and [a1, b1, a1, b1] ⋄ [a2, b2, a2, b2] = { [a, b, a, b]; a = a1 · a2, z ≦P b } , z = min{a1 · b2, +b1, a2, ·b1 + b2, }. Due to Lemma 1 it is evident that (∆K, ⋆) is the commutative weakly associative hypergroup—simply abelian Hν-group. To investigate the associativity law in more detail let us prove the weak associativity law using the concrete form of the structure (K, ∗). We want to verify that ( [ [a, b], [a, b] ] ⋆ [ [c, d], [c, d] ] ) ⋆ [ [p, q], [p, q] ] ∩ [ [a, b], [a, b] ] ⋆ ( [ [c, d], [c, d] ] ⋆ [ [p, q], [p, q] ] ) 6= ∅. Abelization of join spaces 61 for any [a, b], [c, d], [p, q] ∈ K. For the next computation let us denote A = ( [ [a, b], [a, b] ] ⋆ [ [c, d], [c, d] ] ) ⋆ [ [p, q], [p, q] ] , B = [ [a, b], [a, b] ] ⋆ ( [ [c, d], [c, d] ] ⋆ [ [p, q], [p, q] ] ) , z1 = min{a.d + b; c.b + d}, z2 = min{c.q + d; p.d + q}. We get A = { [ [x, y], [x, y] ] ; x = a.c ∧ z1 ≦P y } ⋆ [ [p, q], [p, q] ] = = ⋃ z 1 ≦ P y [ [a.c, y], [a.c, y] ] ⋆ [ [p, q], [p, q] ] = { [ [u, v], [u, v] ] ; u = a.c.p ∧ w1 ≦P v } , where w1 = min{a.c.q + y, p.y + q}. As p ∈ F, p > P 0 F , we have w1 = min { a.c.q + y, p. min{a.d + b, c.b + d} + q } = = min { a.c.q + min{a.d + b, c.b + d}, min{p.a.d + p.b, p.c.b + p.d} + q } = = min { min{a.c.q + a.d + b, a.c.q + c.b + d}, min{p.a.d + p.b + q, p.c.b + p.d + q} } = = min{a.c.q + a.d + b, a.c.q + c.b + d, p.a.d + p.b + q, p.c.b + p.d + q}. So, the set A= { [ [u, v], [u, v] ] ; u = a.c.p ∧ w1 ≦P v } . On the other hand B = [ [a, b], [a, b] ] ⋆ { [ [x, y], [x, y] ] ; x = c.p ∧ z2 ≦P y } = ⋃ z 2 ≦ P y [ [a, b], [a, b] ] ⋆ [ [x, y], [x, y] ] = { [ [u, v], [u, v] ] ; u = a.c.p ∧ w2 ≦P v } , where w2 = min{a.y + b, x.b + y}. Similarly, because a ∈ F, a > P 0F w2 = min { a. min{c.q + d, p.d + q} + b, x.b + min{c.q + d; p.d + q} } = = min { min{a.c.q + a.d, a.p.d + a.q} + b, min{x.b + c.q + d, x.b + p.d + q} } = = min { min{a.c.q + a.d + b, a.p.d + a.q + b}, min{c.p.b + c.q + d, c.p.b + p.d + q} } = = min{a.c.q + a.d + b, a.p.d + a.q + b, c.p.b + c.q + d, c.p.b + p.d + q}. So, the set B= { [ [u, v], [u, v] ] ; u = a.c.p ∧ w2 ≦P v } . Choose [ [u0, v0], [u0, v0] ] ∈ ∆ K , such that u0 = a.c.p and v0 ≧P max{w1, w2}. Evidently, this pair of pairs belongs to A ∩ B. � 62 Š. Hošková Remark 1. Using the previous calculations it is easy to check that the structure (∆ K , ⋆) is never associative. To see it let us first consider F = Q. Necessarily, Q+ ⊂ P. Choose, for example a = 1, b = 2, c = 3, d = −4, p = 5, q = −7. From the above notations we have w1 = min{−23, −19, −17, 3} = −23, w2 = min{−23, −25, 5, 3} = −25. Thus the triple [a, b], [c, d], [p, q] fulfils only the weak associativity law. As it was mentioned earlier each ordered field F contains in itself a copy of the set Q. Thus each time it is possible to find a triple [a, b], [c, d], [p, q] ∈ Q+ × Q such that w1 6= w2 and therefore the structure (∆K, ⋆) is only weakly associative. Remark 2. In the sense of the paper [14] it is possible to define a proximity on this structure, for example, in this way. Let A, B ⊂ K, then A p B if and only if [A)≦ ∩ [B)≦ 6= ∅. where for M ⊂ K we define [M)≦ = ⋃ m∈M [m)≦. We mean the proximity in the sense of the Čech monograph [2]: A relation p on the family of all subsets of the set H is called a proximity on the set H if p satisfies the following conditions: P1. ∅ non p H P2. The relation p is symmetric, i.e., A, B ⊂ H, A p B implies B p A . P3. For any pair of subset A, B ⊂ H, A ∩ B 6= ∅ implies A p B. P4. If A, B, C are subsets of H then (A∪B) p C if and only if either A p C or B p C. Recall that a triad (H, ·, p H ) such that (H, ·) is a hypergroupoid and (H, p H ) is a proximity space will be called a hypergroupoid with a proximity. If for any triad of elements x, y, z ∈ H ( x · (y · z) ) p H ( (x · y) · z ) is valid, then the hyperoperation “·” is called proximally weakly associative–see e.g. [9],[14]. In fact, axioms P 1, P 2 and P 3 from the definition of proximity are obvious. It remains to show that the axiom P 4 is satisfied too. First let us prove the equality [A∪B)≦ = [A)≦ ∪[B)≦, which will be helpful in the next calculations. 1. x ∈ L ⇒ ∃ u ∈ A ∪ B : u ≦ x ⇒ ∃ u ∈ A: u ≦ x or ∃ u ∈ B : u ≦ x ⇒ x ∈ [A)≦ or x ∈ [B)≦ 2. x ∈ P ⇒ x ∈ [A)≦ or x ∈ [B)≦ ⇒ x ∈ [A ∪ B)≦ (By L we mean the left hand side of the equation and by P the right hand side.) Now we can prove the axiom P 4: If A, B, C are subsets of K, then (A∪B) p C if and only if either A p C or B p C. Abelization of join spaces 63 “⇒” (A ∪ B) p C ⇒ [A ∪ B)≦ ∩ [C)≦ 6= ∅ ⇒ ( [A)≦ ∪ [B)≦ ) ∩ [C)≦ 6= ∅ ⇒ A p C or B p C “⇐” A p C or B p C ⇒ [A)≦ ∩[C)≦ 6= ∅ or [B)≦ ∩[C)≦ 6= ∅. Since [A)≦ ⊆ [A∪B)≦ and [B)≦ ⊆ [A ∪ B)≦, we have [A ∪ B)≦ ∩ [C)≦ 6= ∅, i.e., (A ∩ B) p C. Thus we obtain Theorem 2. The hypergroup (K, ∗, p) is a hypergroup with proximity. Theorem 3. The hypergroupoid (∆ K , ⋆) is a weakly associative and commu- tative transposition hypergroup, i.e., a weakly associative join space. Proof. First we will show that the reproduction axiom is fulfilled. Due to Theorem 2 the structure (K, ∗) is the hypergroup, thus [a, b]∗K = K = K∗[a, b] for any [a, b] ∈ K. If we define [a, b] ⋆© [c, d] = { [x, y]; [x, y] ∈ ([a, b] ∗ [c, d]) ∪ ([c, d] ∗ [a, b]) } = { [x, y]; [x, y] ∈ {[a.c, u]; a.d + b ≦P u} ∪ {[c.a, v]; c.b + d ≦P v} } = { [x, y]; x = a.c ∧ z ≦ y) } , where z = min{a.d+b; c.b+d}, evidently [a, b] ⋆©[c, d] ⊃ [a, b]∗[c, d]. Therefore K ⊃ [a, b] ⋆© K ⊃ [a, b] ∗ K = K, which implies that K ⋆© [a, b] = K and similarly [a, b] ⋆© K = K. From this we obtain that reproduction axiom holds in (∆ K , ⋆). Second we will verify the transposition axiom. With respect to the definition of join space and (0.1) we get A = [ [b1, b2], [b1, b2] ]∖[ [a1, a2], [a1, a2] ] = {[ [x, y], [x, y] ] ; [a1, a2] ∈ [b1, b2] ⋆© [x, y] } = {[ [x, y], [x, y] ] ; a1 = b1, .x ∧ a2 ≧P min{b1.y + b2b2.x + y} } = = {[ [x, y], [x, y] ] ; x = a1.b −1 1 ∧ y ≦P max{(a2 − b2).b −1 1 , a2 − b2.a1.b −1 1 } } , B = [ [c1, c2], [c1, c2] ]/[ [d1, d2], [d1, d2] ] = {[ [x, y], [x, y] ] ; [c1, c2] ∈ [x, y] ⋆© [d1, d2] } = {[ [x, y], [x, y] ] ; c1 = d1.x ∧ c2 ≧P min{d1.y + d2, d2.x + y} } = = {[ [x, y], [x, y] ] ; x = c1.d −1 1 ∧ y ≦P max{(c2 − d2).d −1 1 , c2 − d2.c1.d −1 1 } } , C = [ [a1, a2], [a1, a2] ] ⋆ [ [d1, d2], [d1, d2] ] = = {[ [u, v], [u, v] ] ; u = a1.d1 ∧ v ≧P min{a1.d2 + a2, a2.d1 + d2} } , D = [ [b1, b2], [b1, b2] ] ⋆ [ [c1, c2], [c1, c2] ] = = {[ [u, v], [u, v] ] ; u = b1.c1 ∧ v ≧P min{b1.c2 + b2, b2.c1 + c2} } . 64 Š. Hošková We have [ [x, y], [x, y] ] ∈ A ∩ B, i.e., A ≈ B, if and only if x = a1.b −1 1 = c1.d −1 1 and y ≦P min { max{(a2 − b2).b −1 1 , a2 − b2.a1.b −1 1 }, max{(c2 − d2).d −1 1 , c2 − d2.c1.d −1 1 } } . If A ≈ B, then necessarily a1.d1 = b1.c1. Let us denote u0 = a1.d1. For any v0 such that v0 ≧P max { min{a1.d2 + a2, a2.d1 + d2}, min{b1.c2 + b2, b2.c1 + c2} } we obtain [ [u0, v0], [u0, v0] ] ∈ C ∩ D which proves the transposition axiom. � Remark 3. It is easy to verify that under the assumption of the previous theorem even the following equivalence holds: b\a ≈ c/d if and only if a . d ≈ b . c . In fact, one implication follows from Theorem 2. To obtain the converse one (using the notation from the proof of the mentioned theorem) suppose C ≈ D. Thus a1.d1 = b1.c1. Let us denote x0 = a1.b −1 1 = c1.d −1 1 . Then for an arbitrary y0 such that y0 ≦P min { max{(a2 − b2).b −1 1 , a2 − b2.a1.b −1 1 }, max{(c2 − d2).d −1 1 , c2 − d2.c1.d −1 1 } } . we have [ [x0, y0], [x0, y0] ] ∈ A ∩ B, i.e., A ≈ B. References [1] J. Beránek, J. Chvalina, Noncommutative Join Hypergroups of Affine Transformations of Ordered Fields, Dept.Math.Report Series 10, Univ.of South Bohemia, (2002), 15–22. [2] E. Čech, Topological Spaces, revised by Z. Froĺık and M. Katětov, (Academia, Praha 1966). [3] P. Corsini, Prolegomena of Hypergroup Theory, (Aviani Editore Tricesimo 1993). [4] P. Corsini, V. Leoreanu, Applications of Hypergroup Theory, ( Kluwer Academic Pub- lishers, Dordrecht, Hardbound, 2003). [5] J. Chvalina, Functional Graphs, Quasi-ordered Sets and Commutative Hypergroups, MU Brno (1995), (in Czech). [6] J. Chvalina, L. Chvalinová, Transposition hypergroups formed by transformation op- erators on rings of differentiable functions, Ital. J. Pure and Appl. Math., 13 pp. In print. [7] J. Chvalina, Š. Hošková, Abelization of weakly associative hyperstructures based on their direct squares, Acta Mathematica et Informatica Universitatis Ostraviensis, Volume 11/2003, No. 1, 11–25. [8] J. Chvalina, Š. Hošková, Join space of first-order linear partial differential operators with compatible proximity induced by a congruence on their group, Proc. of Mathematical and Computer modelling in Science and Engineering, Prague (2003), 166-170. [9] J. Chvalina, Š. Hošková, Modelling of join spaces with proximities by first-order linear partial differential operators, submitted to Applications of Mathematics, 13p. Abelization of join spaces 65 [10] W. H. Gottschalk, G. A. Hedlund, Topological dynamics, AMS Colloq. Publ. Vol.XXXVI, Providence, Rhode Island, (1974) USA. [11] Š. Hošková, Examples of abelization of hypergroups based on their direct products, Sborńık VA, part B, 2, 7–19, (2002). [12] Š. Hošková, Abelization of differential rings, Proc. of the 1st International mathematical workshop FAST VUT Brno, 2p., (2002). [13] Š. Hošková, Abelization of a certain representation of non-commutative join space, Proc. of International Conference Aplimat 2003, Bratislava, Slovakia, (2003), 365–368. [14] Š. Hošková, J. Chvalina, Abelization of proximal Hν-rings using graphs of good ho- momorphisms and diagonals of direct squares of hyperstructures, Proceedings of 8th Internat. Congress on AHA, Samothraki, Greece (2002), 147–159. [15] J. Jantosciak, Transposition in hypergroups, Algebraic Hyperstructures and Appl. Proc. Sixth Internat. Congress Prague 1996, Democritus Univ. of Thrace Press, Alexandroupo- lis (1997), 77–84. [16] J. Jantosciak, Transposition hypergroups: Noncommutative join spaces, J. Algebra 187 (1997), 97–119. [17] V. Leoreanu, On the hart of join spaces and of regular hypergroups, Riv. Mat. Pura Appl. No. 17 (1995), 133–142. [18] T. Vougiouklis, Hyperstructures and Their Representations, Hadronic Press Monographs in Mathematics, (Palm Harbor Florida 1994). Received September 2003 Accepted January 2005 Šárka Hošková (sarka.hoskova@seznam.cz) Department of Mathematic, University of Defence Brno, Kounicova 65, 612 00 Brno, Czech Republic.