On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point

In this paper we study the semigroup IC(I,[a]) (IO(I,[a])) of closed (open) connected partial homeomorphisms of the unit interval I with a fixed point a∈I. We describe left and right ideals of IC(I,[0]) and the Green's relations on IC(I,[0]). We show that the semigroup IC(I,[0]) is bisimple and...

Повний опис

Збережено в:
Бібліографічні деталі
Дата:2011
Автор: Chuchman, I.
Формат: Стаття
Мова:English
Опубліковано: Інститут прикладної математики і механіки НАН України 2011
Назва видання:Algebra and Discrete Mathematics
Онлайн доступ:http://dspace.nbuv.gov.ua/handle/123456789/154866
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point / I. Chuchman // Algebra and Discrete Mathematics. — 2011. — Vol. 12, № 2. — С. 38–52. — Бібліогр.: 10 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id irk-123456789-154866
record_format dspace
spelling irk-123456789-1548662019-06-17T01:31:06Z On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point Chuchman, I. In this paper we study the semigroup IC(I,[a]) (IO(I,[a])) of closed (open) connected partial homeomorphisms of the unit interval I with a fixed point a∈I. We describe left and right ideals of IC(I,[0]) and the Green's relations on IC(I,[0]). We show that the semigroup IC(I,[0]) is bisimple and every non-trivial congruence on IC(I,[0]) is a group congruence. Also we prove that the semigroup IC(I,[0]) is isomorphic to the semigroup IO(I,[0]) and describe the structure of a semigroup II(I,[0])=IC(I,[0])⊔IO(I,[0]). As a corollary we get structures of semigroups IC(I,[a]) and IO(I,[a]) for an interior point a∈I. 2011 Article On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point / I. Chuchman // Algebra and Discrete Mathematics. — 2011. — Vol. 12, № 2. — С. 38–52. — Бібліогр.: 10 назв. — англ. 1726-3255 2010 Mathematics Subject Classification:20M20,54H15, 20M18. http://dspace.nbuv.gov.ua/handle/123456789/154866 en Algebra and Discrete Mathematics Інститут прикладної математики і механіки НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
description In this paper we study the semigroup IC(I,[a]) (IO(I,[a])) of closed (open) connected partial homeomorphisms of the unit interval I with a fixed point a∈I. We describe left and right ideals of IC(I,[0]) and the Green's relations on IC(I,[0]). We show that the semigroup IC(I,[0]) is bisimple and every non-trivial congruence on IC(I,[0]) is a group congruence. Also we prove that the semigroup IC(I,[0]) is isomorphic to the semigroup IO(I,[0]) and describe the structure of a semigroup II(I,[0])=IC(I,[0])⊔IO(I,[0]). As a corollary we get structures of semigroups IC(I,[a]) and IO(I,[a]) for an interior point a∈I.
format Article
author Chuchman, I.
spellingShingle Chuchman, I.
On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
Algebra and Discrete Mathematics
author_facet Chuchman, I.
author_sort Chuchman, I.
title On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
title_short On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
title_full On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
title_fullStr On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
title_full_unstemmed On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
title_sort on a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point
publisher Інститут прикладної математики і механіки НАН України
publishDate 2011
url http://dspace.nbuv.gov.ua/handle/123456789/154866
citation_txt On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point / I. Chuchman // Algebra and Discrete Mathematics. — 2011. — Vol. 12, № 2. — С. 38–52. — Бібліогр.: 10 назв. — англ.
series Algebra and Discrete Mathematics
work_keys_str_mv AT chuchmani onasemigroupofclosedconnectedpartialhomeomorphismsoftheunitintervalwithafixedpoint
first_indexed 2025-07-14T06:56:05Z
last_indexed 2025-07-14T06:56:05Z
_version_ 1837604448667435008
fulltext Algebra and Discrete Mathematics RESEARCH ARTICLE Volume 12 (2011). Number 2. pp. 38 – 52 c© Journal “Algebra and Discrete Mathematics” On a semigroup of closed connected partial homeomorphisms of the unit interval with a fixed point Ivan Chuchman Communicated by M. Ya. Komarnytskyj Abstract. In this paper we study the semigroup IC(I, [a]) (IO(I, [a])) of closed (open) connected partial homeomorphisms of the unit interval I with a fixed point a ∈ I. We describe left and right ideals of IC(I, [0]) and the Green’s relations on IC(I, [0]). We show that the semigroup IC(I, [0]) is bisimple and every non- trivial congruence on IC(I, [0]) is a group congruence. Also we prove that the semigroup IC(I, [0]) is isomorphic to the semigroup IO(I, [0]) and describe the structure of a semigroup II(I, [0]) = IC(I, [0])⊔IO(I, [0]). As a corollary we get structures of semigroups IC(I, [a]) and IO(I, [a]) for an interior point a ∈ I. 1. Introduction and preliminaries Furthermore we shall follow the terminology of [2] and [6]. For a semigroup S we denote the semigroup S with the adjoined unit by S1 (see [2]). A semigroup S is called inverse if for any element x ∈ S there exists a unique element x−1 ∈ S (called the inverse of x) such that xx−1x = x and x−1xx−1 = x−1. If S is an inverse semigroup, then the function inv : S → S which assigns to every element x of S its inverse element x−1 is called inversion. 2010 Mathematics Subject Classification: 20M20,54H15, 20M18. Key words and phrases: Semigroup of bijective partial transformations, sym- metric inverse semigroup, semigroup of homeomorphisms, group congruence, bisimple semigroup. I . Chuchman 39 If S is a semigroup, then we shall denote the subset of idempotents in S by E(S). If S is an inverse semigroup, then E(S) is closed under multiplication and we shall refer to E(S) a band (or the band of S). If the band E(S) is a non-empty subset of S, then the semigroup operation on S determines the following partial order 6 on E(S): e 6 f if and only if ef = fe = e. This order is called the natural partial order on E(S). A semilattice is a commutative semigroup of idempotents. A semilattice E is called linearly ordered or a chain if its natural order is a linear order. Let E be a semilattice and e ∈ E. We denote ↓e = {f ∈ E | f 6 e} and ↑e = {f ∈ E | e 6 f}. If S is a semigroup, then we shall denote by R, L , J , D and H the Green relations on S (see [2]): aRb if and only if aS1 = bS1; aL b if and only if S1a = S1b; aJ b if and only if S1aS1 = S1bS1; D = L ◦ R = R ◦ L ; H = L ∩ R. A semigroup S is called simple if S does not contain proper two-sided ideals and bisimple if S has only one D-class. A congruence C on a semigroup S is called non-trivial if C is distinct from universal and identity congruence on S, and group if the quotient semigroup S/C is a group. The bicyclic semigroup C (p, q) is the semigroup with the identity 1 generated by elements p and q subject only to the condition pq = 1. The distinct elements of C (p, q) are exhibited in the following useful array: 1 p p2 p3 · · · q qp qp2 qp3 · · · q2 q2p q2p2 q2p3 · · · q3 q3p q3p2 q3p3 · · · ... ... ... ... . . . The bicyclic semigroup is bisimple and every one of its congruences is either trivial or a group congruence. Moreover, every non-annihilating homomorphism h of the bicyclic semigroup is either an isomorphism or the image of C (p, q) under h is a cyclic group (see [2, Corollary 1.32]). The bicyclic semigroup plays an important role in algebraic theory of semigroups and in the theory of topological semigroups. For example the well-known Andersen’s result [1] states that a (0–)simple semigroup 40On a semigroup of closed connected partial homeomorphisms is completely (0–)simple if and only if it does not contain the bicyclic semigroup. Let IX denote the set of all partial one-to-one transformations of an non-empty set X together with the following semigroup operation: x(αβ) = (xα)β if x ∈ dom(αβ) = {y ∈ domα | yα ∈ domβ} , for α, β ∈ IX . The semigroup IX is called the symmetric inverse semi- group over the setX (see [2, Section 1.9]). The symmetric inverse semigroup was introduced by Wagner [10] and it plays a major role in the theory of semigroups. Let I be an interval [0, 1] with the usual topology. A partial map α : I ⇀ I is called: • closed, if domα and ranα are closed subsets in I; • open, if domα and ranα are open subsets in I; • convex, if domα and ranα are convex non-singleton subsets in I; • monotone, if x1 6 x2 implies (x1)α 6 (x2)α, for all x1, x2 ∈ domα; • a local homeomorphism, if the restriction α|domα : domα → ranα is a homeomorphism. We fix an arbitrary a ∈ I. Hereafter we shall denote by: • IC(I, [a]) the semigroup of all closed connected partial homeomor- phisms α such that IntI(domα) 6= ∅ and (a)α = a; • IO(I, [a]) the semigroup of all open connected partial homeomor- phisms α such that (a)α = a; • H(I) the group of all homeomorphisms of I; • H ր (I) the group of all monotone homeomorphisms of I; • I the identity map from I onto I. Remark 1. We observe that for every a ∈ I the semigroups IC(I, [a]) and IO(I, [a]) are inverse subsemigroups of the symmetric inverse semigroup II over the set I. In [3, 4] Gluskin studied the semigroup S of homeomorphic transfor- mations of the unit interval. He described all ideals, homomorphisms and automorphisms of the semigroup S and congruence-free subsemigroups of S. This studies was continued in [7] by Shneperman. In [9] Shneperman I . Chuchman 41 described the structure of the semigroup of homeomorphisms of a sim- ple arc. In the paper [8] he studied a semigroup G(X) of all continuous transformations of a closed subset X of the real line. In our paper we study the semigroup IC(I, [a]) (IO(I, [a])) of closed (open) connected partial homeomorphisms of the unit interval I with a fixed point a ∈ I. We describe left and right ideals of IC(I, [0]) and the Green’s relations on IC(I, [0]). We show that the semigroup IC(I, [0]) is bisimple and every non-trivial congruence on IC(I, [0]) is a group congruence. Also we prove that the semigroup IC(I, [0]) is isomorphic to the semigroup IO(I, [0]) and describe the structure of a semigroup II(I, [0]) = IC(I, [0]) ⊔ IO(I, [0]). As a corollary we get structures of semigroups IC(I, [a]) and IO(I, [a]) for an interior point a ∈ I. 2. On the semigroup IC(I, [0]) Proposition 1. The following conditions hold: (i) every element of the semigroup IC(I, [0]) (IO(I, [1])) is a monotone partial map; (ii) the semigroups IC(I, [0]) and IC(I, [1]) are isomorphic; (iii) max{domα} exists for every α ∈ IC(I, [0]); (iv) sup{domα} exists for every α ∈ IO(I, [0]); (v) (0)α = 0 and (1)α = 1 for every α ∈ H ր (I). Proof. Statements (i), (iii), (iv) and (v) follow from elementary properties of real-valued continuous functions. (ii) A homomorphism i : IC(I, [0]) → IC(I, [1]) we define by the following way: (α)i = β, where domβ = {1− x | x ∈ domα}, ranβ = {1− x | x ∈ ranα}, and (a)β = 1− (1− a)α for all a ∈ domβ. Simple verifications show that such defined map i is an isomorphism from the semigroup IC(I, [0]) onto the semigroup IO(I, [1]). Proposition 2. The following statements hold: (i) an element α of the semigroup IC(I, [0]) is an idempotent if and only if (x)α = x for every x ∈ domα; 42On a semigroup of closed connected partial homeomorphisms (ii) If ε, ι ∈ E(IC(I, [0])), then ε 6 ι if and only if dom ε ⊆ dom ι; (iii) The semilattice E(IC(I, [0])) is isomorphic to the semilattice ((0, 1],min) under the mapping (ε)h = max{dom ε}; (iv) αRβ in IC(I, [0]) if and only if domα = domβ; (v) αL β in IC(I, [0]) if and only if ranα = ranβ. (vi) αH β in IC(I, [0]) if and only if domα = domβ and ranα = ranβ. (vii) for every distinct idempotents ε, ι ∈ IC(I, [0]) there exists an element α of the semigroup IC(I, [0]) such that α · α−1 = ε and α−1 · α = ι; (viii) αDβ for all α, β ∈ IC(I, [0]), and hence the semigroup IC(I, [0]) is bisimple; (ix) αJ β for all α, β ∈ IC(I, [0]), and hence the semigroup IC(I, [0]) is simple; (x) a subset L is a left ideal of IC(I, [0]) if and only if there exists a ∈ (0, 1] such that either L = {α ∈ IC(I, [0]) | ranα ⊆ [0, a)} or L = {α ∈ IC(I, [0]) | ranα ⊆ [0, a]}; (xi) a subset R is a right ideal of IC(I, [0]) if and only if there exists a a ∈ (0, 1] such that either R = {α ∈ IC(I, [0]) | domα ⊆ [0, a)} or R = {α ∈ IC(I, [0]) | domα ⊆ [0, a]}. Proof. Statements (i), (ii) and (iii) are trivial and they follow from the definition of the semigroup IC(I, [0]). (iv) Let be α, β ∈ IC(I, [0]) such that αRβ. Since αIC(I, [0]) = βIC(I, [0]) and IC(I, [0]) is an inverse semigroup, Theorem 1.17 [2] implies that αIC(I, [0]) = αα−1IC(I, [0]) and βIC(I, [0]) = ββ−1IC(I, [0]), and hence we have that αα−1 = ββ−1. Therefore we get that domα = domβ. Conversely, let be α, β ∈ IC(I, [0]) such that domα = domβ. Then αα−1 = ββ−1. Since IC(I, [0]) is an inverse semigroup, Theorem 1.17 [2] implies that αIC(I, [0]) = αα−1IC(I, [0]) = ββ−1IC(I, [0]) = βIC(I, [0]), and hence αIC(I, [0]) = βIC(I, [0]). The proof of statement (v) is similar to (iv). I . Chuchman 43 Statement (vi) follows from (iv) and (v). (vii) We fix arbitrary distinct idempotents ε and ι in IC(I, [0]). If dε = max{dom ε} and dι = max{dom ι}, then dε 6= 0, dι 6= 0, and ε and ι are identity maps of intervals [0, dε] and [0, dι], respectively. We define a partial map α : I ⇀ I as follows: domα = [0, dε], ranα = [0, dι] and (x)α = dι dε ·x, for all x ∈ domα. Then we have that α ∈ IC(I, [0]), α · α−1 = ε and α−1 · α = ι. (viii) Statement (vii) and Lemma 1.1 from [5] imply that IC(I, [0]) is a bisimple semigroup. Since D ⊆ J , statement (viii) implies assertion (ix). (x) The semigroup operation on IC(I, [0]) implies that the sets {α ∈ IC(I, [0]) | ranα ⊆ [0, a)} and {α ∈ IC(I, [0]) | ranα ⊆ [0, a]} are left ideals in IC(I, [0]), for every a ∈ (0, 1]. Suppose that L is an arbitrary left ideal of the semigroup IC(I, [0]). We fix any α ∈ L . Then statements (i), (ii) and (v) imply that the left ideal L contains all β ∈ L such that ranβ ⊆ ranα. We put A = ⋃ α∈L ranα and let a = supA. If there exists α ∈ L such that sup ranα = a then statement (v) implies that L = {α ∈ IC(I, [0]) | ranα ⊆ [0, a]}. In other case we have that statement (v) implies that L = {α ∈ IC(I, [0]) | ranα ⊆ [0, a)} . The proof of statement (xi) is similar to statement (x). Definitions of the group H ր (I) and the semigroup IC(I, [0]) imply the following: Proposition 3. The group of units of the semigroup IC(I, [0]) is isomor- phic to (i.e., coincides with) the group H ր (I). Proposition 2.20 of [2] states that every two subgroup which lie in some D-class are isomorphic, and hence Proposition 3 implies the following: Corollary 1. Every maximal subgroup of the semigroup IC(I, [0]) is isomorphic to H ր (I). Later we need the following two lemmas: 44On a semigroup of closed connected partial homeomorphisms Lemma 1. Let R is an arbitrary congruence on a semilattice E and let a and b be elements of the semilattice E such that aRb. If a 6 b then aRc for all c ∈ E such that a 6 c 6 b. The proof of the lemma follows from the definition of a congruence on a semilattice. Lemma 2. For arbitrary distinct idempotents α and β of the semigroup IC(I, [0]) there exists a subsemigroup C in IC(I, [0]) such that α, β ∈ C and C is isomorphic to the bicyclic semigroup C (p, q). Proof. Without loss of generality we can assume that β 6 α in E(IC(I, [0])). We define partial maps γ, δ : I ⇀ I as follows: dom γ = [0, dα], ran γ = [0, dβ ] and (x)γ = dβ dα ·x, for all x ∈ dom γ, and dom δ = [0, dβ ], ran δ = [0, dα] and (x)δ = dα dβ · x, for all x ∈ dom δ, where dα = max{domα} and dβ = max{domβ}. Then we have that α · γ = γ · α = γ, α · δ = δ · α = δ, γ · δ = α and δ · γ = β 6= α. Hence by Lemma 1.31 from [2] we get that a subsemigroup in IC(I, [0]) which is generated by elements γ and δ is isomorphic to the bicyclic semigroup C (p, q). Theorem 1. Every non-trivial congruence on the semigroup IC(I, [0]) is a group congruence. Proof. Suppose that K is a non-trivial congruence on the semigroup IC(I, [0]). Then there exist distinct elements α and β in IC(I, [0]) such that αKβ. We consider the following three cases: (i) α and β are idempotents in IC(I, [0]); (ii) α and β are not H -equivalent in IC(I, [0]); (iii) α and β are H -equivalent in IC(I, [0]). Suppose case (i) holds and without loss of generality we assume that α 6 β in E(IC(I, [0])). We define a partial map ρ : I ⇀ I as follows: dom ρ = domβ, ran ρ = I and (x)ρ = 1 dβ · x, for all x ∈ dom ρ, I . Chuchman 45 where dβ = max{domβ}. Then we have that ρ−1 · β · ρ = I and hence by Proposition 1(i) the element αβ = ρ−1 · α · ρ is an idempotent of the semigroup IC(I, [0]). Obviously, αβ 6 I in E(IC(I, [0])), αβ 6= I and αβKI. Then by Lemma 2 there exist γ, δ ∈ IC(I, [0]) such that I·γ = γ ·I = γ, I·δ = δ·I = δ, γ ·δ = I and δ·γ = αβ 6= I, and a subsemigroup C 〈γ, δ〉 in IC(I, [0]) which is generated by elements γ and δ is isomorphic to the bicyclic semigroup C (p, q). Since by Corol- lary 1.32 from [2] every non-trivial congruence on the bicyclic semigroup C (p, q) is a group congruence on C (p, q) we get that all idempotents of the semigroup C 〈γ, δ〉 are K-equivalent. Also by Lemma 1.31 from [2] we get that every idempotent of the semigroup C 〈γ, δ〉 has a form δn · γn = (δ · . . . · δ ︸ ︷︷ ︸ n−times ) · (γ · . . . · γ ︸ ︷︷ ︸ n−times ), where n = 0, 1, 2, 3, . . . , and hence we get that dom ( δn · γn ) = [0, dn], where d = max{domαβ}. This implies that for every idempotent ε ∈ IC(I, [0]) there exists a positive integer n such that δn ·γn 6 ε, and hence by Lemma 1 we get that all idem- potents of the semigroup IC(I, [0]) are K-equivalent. Then Lemma 7.34 and Theorem 7.36 from [2] imply that the quotient semigroup IC(I, [0])/K is a group. Suppose case (ii) holds: α and β are not H -equivalent in IC(I, [0]). Since IC(I, [0]) is an inverse semigroup we get that either αα−1 6= ββ−1 or α−1α 6= β−1β. Suppose inequality αα−1 6= ββ−1 holds. Since αKβ and IC(I, [0]) is an inverse semigroup, Lemma III.1.1 from [6] implies that ( αα−1 ) K ( ββ−1 ) , and hence by case (i) we get that K is a group congruence on the semigroup IC(I, [0]). In the case α−1α 6= β−1β the proof is similar. Suppose case (iii) holds: α and β are H -equivalent in IC(I, [0]). Then Theorem 2.3 of [2] implies that without loss of generality we can assume that α and β are elements of the group of units H(I) of the semigroup IC(I, [0]). Therefore we get that I = α · α−1 and γ = β · α−1 ∈ H(I) are H -equivalent distinct elements in IC(I, [0]). Since I 6= γ we get that there exists xγ ∈ I such that (xγ)γ 6= xγ . We suppose (xγ)γ > xγ . We define a partial map δ : I ⇀ I as follows: dom δ = [0, (xγ)γ], ran δ = [0, xγ ] and (x)ρ = xγ (xγ)γ · x, for all x ∈ dom δ. Then we have that I · δ = δ and hence we get that (γ · δ)Kδ. Since dom(γ · δ) = dom γ 6= dom δ, Proposition 1(vi) implies 46On a semigroup of closed connected partial homeomorphisms that the elements γ · δ and δ are not H -equivalent. Therefore case (ii) holds, and hence K is a group congruence on the semigroup IC(I, [0]). In the case (xγ)γ < xγ the proof that K is a group congruence on the semigroup IC(I, [0]) is similar. This completes the proof of our theorem. Proposition 4. The semigroups IC(I, [0]) and IO(I, [0]) are isomorphic. Proof. We define a map i : IC(I, [0]) → IO(I, [0]) by the following way: for arbitrary α ∈ IC(I, [0]) we put (α)i is the restriction of α on the set [0, aα] \ {aα}, where aα = max{domα}, with dom((α)i) = domα \ {aα} and ran((α)i) = ranα \ {(aα)α}. Simple verifications show that such defined map i : IC(I, [0]) → IO(I, [0]) is an isomorphism. 3. On the semigroup II(I, [0]) We put II(I, [0]) = IC(I, [0]) ⊔ IO(I, [0]). Later we shall denote elements of the semigroup IC(I, [0]) by α and put ◦ α = (α)i ∈ IO(I, [0]), where i : IC(I, [0]) → IO(I, [0]) is the isomor- phism which is defined in the proof of Proposition 4. Since the semigroups IC(I, [0]) and IO(I, [0]) are inverse subsemigroups of the symmetric in- verse semigroup II over the set I and by Proposition 1 all elements of the semigroups IC(I, [0]) and IO(I, [0]) are monotone partial maps, the semi- group operation in II implies that for α ∈ IC(I, [0]) and ◦ β ∈ IO(I, [0]) we have that α · ◦ β = { γ, if ranα ⊂ dom ◦ β; ◦ γ, if dom ◦ β ⊂ ranα and ◦ β ·α = { ◦ δ, if ran ◦ β ⊂ domα; δ, if domα ⊂ ran ◦ β, where γ = α · β ∈ IC(I, [0]) (i.e., ◦ γ = ◦ α · ◦ β ∈ IO(I, [0])) and δ = β · α ∈ IC(I, [0]) (i.e., ◦ δ = ◦ β · ◦ α ∈ IO(I, [0])). Hence we get the following: Proposition 5. II(I, [0]) is an inverse semigroup. Given two partially ordered sets (A,6A) and (B,6B), the lexicograph- ical order 6lex on the Cartesian product A×B is defined as follows: (a, b) 6lex (a′, b′) if and only if a <A a′ or (a = a′ and b 6B b′). In this case we shall say that the partially ordered set (A × B,6lex) is the lexicographic product of partially ordered sets (A,6A) and (B,6B) and it is denoted by A×lex B. We observe that a lexicographic order of two linearly ordered sets is a linearly ordered set. I . Chuchman 47 Hereafter for every α ∈ IC(I, [0]) and ◦ β ∈ IO(I, [0]) we denote dα = max{domα}, rα = max{ranα}, dβ = sup{dom ◦ β} and rβ = sup{ran ◦ β}. Obviously we have that dα = sup{dom ◦ α} and rα = sup{ran ◦ α} for any ◦ α ∈ IO(I, [0]). Proposition 6. The following conditions hold: (i) E(II(I, [0])) = E(IC(I, [0])) ∪ E(IO(I, [0])). (ii) If α, ◦ α, β, ◦ β ∈ E(II(I, [0])), then (a) ◦ α 6 α; (b) α 6 β if and only if dα 6 dβ (rα 6 rβ); (c) ◦ α 6 ◦ β if and only if dα 6 dβ (rα 6 rβ); (d) α 6 ◦ β if and only if dα < dβ (rα < rβ); and (e) ◦ α 6 β if and only if dα 6 dβ (rα 6 rβ). (iii) The semilattice E(II(I, [0])) is isomorphic to the lexicographic product (0; 1] ×lex {0; 1} of the semilattices ((0; 1],min) and ({0; 1},min) under the mapping (α)i = (dα; 1) and ( ◦ α)i = (dα; 0), and hence E(II(I, [0])) is a linearly ordered semilattice. (iv) The elements α and β of the semigroup II(I, [0]) are R-equivalent in II(I, [0]) provides either α, β ∈ IC(I, [0]) or α, β ∈ IO(I, [0]) and moreover, we have that (a) αRβ in II(I, [0]) if and only if dα = dβ; and (b) ◦ αR ◦ β in II(I, [0]) if and only if dα = dβ. (v) The elements α and β of the semigroup II(I, [0]) are L -equivalent in II(I, [0]) provides either α, β ∈ IC(I, [0]) or α, β ∈ IO(I, [0]) and moreover, we have that (a) αL β in II(I, [0]) if and only if rα = rβ; and (b) ◦ αL ◦ β in II(I, [0]) if and only if rα = rβ. (vi) The elements α and β of the semigroup II(I, [0]) are H -equivalent in II(I, [0]) provides either α, β ∈ IC(I, [0]) or α, β ∈ IO(I, [0]) and moreover, we have that (a) αH β in II(I, [0]) if and only if dα = dβ and rα = rβ; and (b) ◦ αH ◦ β in II(I, [0]) if and only if dα = dβ and rα = rβ. 48On a semigroup of closed connected partial homeomorphisms (vii) II(I, [0]) is a simple semigroup. (viii) The semigroup II(I, [0]) has only two distinct D-classes: that are inverse subsemigroups IC(I, [0]) and IO(I, [0]). Proof. Statements (i), (ii) and (iii) follow from the definition of the semigroup II(I, [0]) and Proposition 5. Proofs of statements (iv), (v) and (vi) follow from Proposition 5 and Theorem 1.17 [2] and are similar to statements (iv), (v) and (vi) of Proposition 2. (vii) We shall show that II(I, [0]) · α · II(I, [0]) = II(I, [0]) for every α ∈ II(I, [0]). We fix arbitrary α, β ∈ II(I, [0]) and show that there exist γ, δ ∈ II(I, [0]) such that γ · α · δ = β. We consider the following four cases: (1) α = α ∈ IC(I, [0]) and β = β ∈ IC(I, [0]); (2) α = α ∈ IC(I, [0]) and β = ◦ β ∈ IO(I, [0]); (3) α = ◦ α ∈ IO(I, [0]) and β = β ∈ IC(I, [0]); (4) α = ◦ α ∈ IO(I, [0]) and β = ◦ β ∈ IO(I, [0]). By Λb a we denote a linear partial map from I into I with domΛb a = [0; a] and ranΛb a = [0; b], and defined by the formula: (x)Λb a = b a · x, for x ∈ domΛb a. We put: γ = Λdα dβ and δ = α−1 · Λ dβ dα · β in case (1); γ = Λdα dβ and δ = α−1 · Λ dβ dα · β in case (2); γ = Λa dβ and δ = α−1 · Λ dβ a · β, where 0 < a < dα, in case (3); γ = Λdα dβ and δ = α−1 · Λ dβ dα · β in case (4). Elementary verifications show that γ · α · δ = β, and this completes the proof of assertion (vii). Statement (viii) follows from statements (iv) and (v). On the semigroup II(I, [0]) we determine a relation ∼id by the follow- ing way. Let i : IC(I, [0]) → IO(I, [0]) be a map which is defined in the proof of Proposition 4. We put α ∼id β if and only if α = β or (α)id = β or (β)id = α, I . Chuchman 49 for α, β ∈ II(I, [0]). Simple verifications show that ∼id is an equivalence relation on the semigroup II(I, [0]). The following proposition immediately follows from Proposition 1(i) and the definition of the relation ∼id on the semigroup II(I, [0]): Proposition 7. Let α and β are elements of the semigroup II(I, [0]). Then α ∼id β in II(I, [0]) if and only if the following conditions hold: (i) dα = dβ; (ii) rα = rβ; (iii) (x)α = (x)β for every x ∈ [0, dα); (iv) (y)α = (y)β for every y ∈ [0, dβ). Proposition 8. The relation ∼id is a congruence on the semigroup II(I, [0]). Moreover, the quotient semigroup II(I, [0])/∼id is isomorphic to the semigroup IC(I, [0]). Proof. We fix arbitrary α, ◦ α, β, ◦ γ ∈ II(I, [0]). It is complete to show that the following conditions hold: (i) ( α · β ) ∼id ( ◦ α · β ) ; (ii) ( β · α ) ∼id ( β · ◦ α ) ; (iii) ( α · ◦ γ ) ∼id ( ◦ α · ◦ γ ) ; (iv) ( ◦ γ · α ) ∼id ( ◦ γ · ◦ α ) . Suppose case (i) holds. If dβ 6 rα, then Proposition 1(i) implies that (x) ( α · β ) = (x) ( ◦ α · β ) for all x ∈ [ 0, (dβ)(α) −1 ) , and hence by Proposi- tion 7 we get that ( α · β ) ∼id ( ◦ α · β ) . If dβ > rα, then Proposition 1(i) implies that (x) ( α · β ) = (x) ( ◦ α · β ) for all x ∈ [0, dα), and hence by Proposition 7 we get that ( α · β ) ∼id ( ◦ α · β ) . In cases (ii), (iii) and (iv) the proofs are similar. Hence ∼id is a congruence on the semigroup II(I, [0]). Let Φ∼id : II(I, [0]) → IC(I, [0]) a natural homomorphism which is generated by the congruence ∼id. Since the restriction Φ∼id |IC(I,[0]) : IC(I, [0]) → IC(I, [0]) of the natural homomorphism Φ∼id : II(I, [0]) → IC(I, [0]) is an identity map we conclude that the semigroup (II(I, [0]))Φ∼id is isomorphic to the semigroup IC(I, [0]). Theorem 2. Let K be a non-trivial congruence on the semigroup II(I, [0]). Then the quotient semigroup II(I, [0])/K is either a group or II(I, [0])/K is isomorphic to the semigroup IC(I, [0]). 50On a semigroup of closed connected partial homeomorphisms Proof. Since the subsemigroup of idempotents of the semigroup II(I, [0]) is linearly ordered we have that similar arguments as in the proof of Theorem 1 imply that there exist distinct idempotents ε and ι in II(I, [0]) such that εKι and ε 6 ι. If the set (ε, ι) = {υ ∈ E(II(I, [0])) | ε < υ < ι} is non-empty, then Lemma 1 and Theorem 1 imply that the quotient semigroup II(I, [0])/R is inverse and it contains only one idempotent, and hence by Lemma II.1.10 from [6] we get that II(I, [0])/R is a group. Otherwise Proposition 7(ii) implies that ε = ◦ α and ι = α for some idempotents ◦ α ∈ IO(I, [0]) and α ∈ IC(I, [0]). Since by Proposition 2(ix) the semigroup IC(I, [0]) is simple we get that for every β ∈ IC(I, [0]) there exist γ, δ ∈ IC(I, [0]) such that β = γ · α · δ. Since IC(I, [0]) is an inverse semigroup and all elements of IC(I, [0]) are monotone partial maps of the unit interval I we conclude that that β = β · β −1 · γ · α · α−1 · α · α−1 · α · δ · β −1 · β, and hence for elements γβ = β · β −1 · γ · α · α−1 and δβ = α−1 · α · δ · β −1 · β, of the semigroup IC(I, [0]) the following conditions hold: β = γβ · α · δβ , domβ = dom γβ , ran γβ = domα, ranα = dom δβ and ranβ = ran δβ . Analogously, since all elements of the semigroups IC(I, [0]) and IO(I, [0]) are monotone partial maps of I we get that ◦ β = ◦ γβ · ◦ α · ◦ δβ and hence βR ◦ β. This implies that the congruence R on the semigroup II(I, [0]) coincides with the congruence ∼id on II(I, [0]). Then Proposition 8 implies that the quotient semigroup II(I, [0])/K is isomorphic to the semigroup IC(I, [0]). By S2 we denote the cyclic group of order 2. Theorem 3. For arbitrary a, b ∈ (0, 1) the semigroups IC(I, [a]) and IC(I, [b]) are isomorphic. Moreover, for every a ∈ (0, 1) the semigroup IC(I, [a]) is isomorphic to the direct product S2 × IC(I, [0])× IC(I, [0]). Proof. We fix an arbitrary a ∈ (0, 1). Obviously, the semigroup IC(I, [a]) is isomorphic to the direct product S2 × IC ր (I, [a]), where IC ր (I, [a]) is a subsemigroup of IC(I, [a]) which consists of monotone partial maps of the unit interval I. I . Chuchman 51 By IC ր (I ⊔ I, [0]) we denote the semigroup of all monotone convex closed partial local homeomorphisms α of the interval [−1, 1] such that (0)α = 0 and 0 ∈ Int[−1,1](domα). We define a map i : IC ր (I, [a]) → IC ր (I ⊔ I, [0]) by the following way. For an arbitrary α ∈ IC ր (I, [a]) we determine a partial map β = (α)i ∈ IC ր (I ⊔ I, [0]) as follows: (i) domβ = [ dm(α)− a a , dM (α)− a 1− a ] , where dm(α) = min{domα} and dM (α) = max{domα}; (ii) ranβ = [ rm(α)− a a , rM (α)− a 1− a ] , where rm(α) = min{ranα} and rM (α) = max{ranα}; and (iii) (x)β = { (ax+ a)α, if x 6 0 ((1− a)x+ a)α, if x > 0 , for all x ∈ domβ. Simple verifications show that such defined map i : IC ր (I, [a]) → IC ր (I ⊔ I, [0]) is an isomorphism. This completes the first part of the proof of the theorem. Next we define a map j : IC ր (I ⊔ I, [0]) → IC(I, [0]) × IC(I, [0]) by the following way. For an arbitrary α ∈ IC ր (I ⊔ I, [0]) we determine a pair of partial maps (β, γ) = (α)i ∈ IC(I, [0])× IC(I, [0]) as follows: (i) domβ = domα ∩ [0, 1] and ranβ = ranα ∩ [0, 1]; (ii) dom γ = {−x | x ∈ domα ∩ [0, 1]} and ran γ = {−x | x ∈ ranα ∩ [0, 1]}; (iii) (x)β = (x)α for x ∈ domβ; and (iv) (x)γ = −(x)α for x ∈ dom γ. Simple verifications show that such defined map j : IC ր (I ⊔ I, [0]) → IC(I, [0])× IC(I, [0]) is an isomorphism. This completes the proof of the theorem. Theorem 3 implies the following: Corollary 2. For arbitrary a, b ∈ (0, 1) the semigroups II(I, [a]) and II(I, [b]) are isomorphic. Moreover, for every a ∈ (0, 1) the semigroup II(I, [a]) is isomorphic to the direct product S2 × II(I, [0])× II(I, [0]). 52On a semigroup of closed connected partial homeomorphisms References [1] O. Andersen, Ein Bericht über die Struktur abstrakter Halbgruppen, PhD Thesis, Hamburg, 1952. [2] A. H. Clifford and G. B. Preston, The Algebraic Theory of Semigroups, Vol. I., Amer. Math. Soc. Surveys 7, Providence, R.I., 1961; Vol. II., Amer. Math. Soc. Surveys 7, Providence, R.I., 1967. [3] L. M. Gluskin, The semi-group of homeomorphic mappings of an interval, Mat. Sb. (N.S.), 49(91):1 (1959), 13—28 (in Russian). [4] L. M. Gluskin, On some semigroup of continuous functions, Dopovidi AN URSR no. 5 (1960), 582—585 (in Ukrainian). [5] W. D. Munn, Uniform semilattices and bisimple inverse semigroups, Quart. J. Math. 17:1 (1966), 151—159. [6] M. Petrich, Inverse Semigroups, John Wiley & Sons, New York, 1984. [7] L. B. Shneperman, Semigroups of continuous transformations, Dokl. Akad. Nauk SSSR 144:3 (1962), 509—511 (in Russian). [8] L. B. Shneperman, Semigroups of continuous transformations of closed sets of the number axis, Izv. Vyssh. Uchebn. Zaved. Mat., no. 6 (1965), 166—175 (in Russian). [9] L. B. Shneperman, The semigroups of the homeomorphisms of a simple arc, Izv. Vyssh. Uchebn. Zaved. Mat., no. 2 (1966), 127—136 (in Russian). [10] V. V. Wagner, Generalized groups, Dokl. Akad. Nauk SSSR 84 (1952), 1119—1122 (in Russian). Contact information Ivan Chuchman Department of Mechanics and Mathematics, Ivan Franko Lviv National University, Universytetska 1, Lviv, 79000, Ukraine E-Mail: chuchman i@mail.ru Received by the editors: 22.09.2011 and in final form 22.09.2011.