About this Journal Submit a Manuscript Table of Contents
ISRN Geometry
Volume 2012 (2012), Article ID 359403, 9 pages
doi:10.5402/2012/359403
Research Article

Homotopy Extension Property in Homotopy Theory for Topological Semigroups

Department of Mathematics and Institute for Mathematical Research, University Putra Malaysia, 43400 UPM, Serdang, Selangor, Malaysia

Received 2 November 2011; Accepted 13 December 2011

Academic Editor: D. Danielli

Copyright © 2012 Adem Kılıçman and Amin Saif. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The purpose of this paper is to extend the concept of homotopy extension property in homotopy theory for topological spaces to its analogical structure in homotopy theory for topological semigroups. In this extension, we also give some results concerning on absolutely retract and its properties.

1. Introduction

The definition of homotopy theory for topological semigroups and most of the backgrounds for this paper have been worked out previously by Čerin in [1]. He introduced the concepts of 𝑆 -homotopy relation, pathwise 𝑆 -connectedness, 𝑆 -homotopy domination, 𝑆 -contractibility, and 𝑆 -fibration.

A topological semigroup ( 𝐴 , ) is called an 𝑆 -subspace of a topological semigroup ( 𝑆 , ) if 𝐴 is a subspace of 𝑆 , the map takes the product 𝐴 × 𝐴 into 𝐴 and ( 𝑥 , 𝑦 ) = ( 𝑥 , 𝑦 ) for all 𝑥 , 𝑦 𝐴 . The pair ( 𝑆 𝐼 , 𝑝 ( ) ) is a topological semigroup with the compact-open topology on a path space 𝑆 𝐼 , where the multiplication 𝑝 ( ) 𝑆 𝐼 × 𝑆 𝐼 𝑆 𝐼 on 𝑃 𝑎 ( 𝑆 ) defined by [ ] 𝑝 ( ) ( 𝛼 , 𝛽 ) ( 𝑡 ) = ( 𝛼 ( 𝑡 ) , 𝛽 ( 𝑡 ) ) , f o r 𝛼 , 𝛽 𝑆 𝐼 , 𝑡 𝐼 . ( 1 . 1 )

𝑃 ( 𝑆 , ) will denote a topological semigroup ( 𝑆 𝐼 , 𝑝 ( ) ) . For 𝑠 𝑆 , ̃ 𝑠 will denote the constant path into 𝑠 in ( 𝑆 , ) . The function 𝑓 ( 𝑆 1 , ) ( 𝑆 2 , ) is called an 𝑆 -map of ( 𝑆 1 , ) into ( 𝑆 2 , ) if 𝑓 is a continuous map of space 𝑆 1 into 𝑆 2 such that 𝑓 ( ( 𝑥 , 𝑦 ) ) = ( 𝑓 ( 𝑥 ) , 𝑓 ( 𝑦 ) ) , f o r 𝑥 , 𝑦 𝑆 1 . ( 1 . 2 )

Two 𝑆 -maps 𝑓 , 𝑔 ( 𝑆 , ) ( 𝑂 , ) are called 𝑆 -homotopic (written 𝑓 𝑠 𝑔 ) if there is an 𝑆 -map 𝐻 ( 𝑆 , ) 𝑃 ( 𝑂 , ) (called 𝑆 -homotopy) such that 𝐻 ( 𝑠 ) ( 0 ) = 𝑓 ( 𝑠 ) and 𝐻 ( 𝑠 ) ( 1 ) = 𝑔 ( 𝑠 ) for all 𝑠 𝑆 , for more details see [13].

Recall [1] that a surjective 𝑆 -map 𝑓 ( 𝑆 , ) ( 𝑂 , ) is called 𝑆 -fibration if it has an 𝑆 -homotopy lifting property for any element in , where is the collection of all topological semigroup. Now we introduce the concept of an 𝑆 -lifting function of an 𝑆 -fibration as follows: let 𝑓 ( 𝑆 , ) ( 𝑂 , ) be an 𝑆 -map. Then the function 𝐿 𝑓 ( Δ 𝑓 , × 𝑝 ( ) ) 𝑃 ( 𝑆 , ) is called an 𝑆 -lifting function for 𝑓 if it is satisfying the following:(1) 𝐿 𝑓 is an 𝑆 -map,(2) 𝐿 𝑓 ( 𝑠 , 𝛼 ) ( 0 ) = 𝑠 for ( 𝑠 , 𝛼 ) Δ 𝑓 ,(3) 𝑓 [ 𝐿 𝑓 ( 𝑠 , 𝛼 ) ] = 𝛼 for ( 𝑠 , 𝛼 ) Δ 𝑓 ,

where Δ 𝑓 = { ( 𝑠 , 𝛼 ) 𝑆 × 𝑃 𝑎 ( 𝑂 ) 𝑓 ( 𝑠 ) = 𝛼 ( 0 ) } . If 𝐿 𝑓 ( 𝑠 , 𝑓 ̃ 𝑠 ) = ̃ 𝑠 for all 𝑠 𝑆 , then the 𝑆 -lifting function is called an 𝑆 -regular lifting function. Also we say that the 𝑆 -map 𝑓 ( 𝑆 , ) ( 𝑂 , ) is an 𝑆 -regular fibration if it has 𝑆 -regular lifting function.

Note that we can easily prove that 𝑆 -map 𝑓 is an 𝑆 -regular fibration if and only if it has 𝑆 -regular lifting function. Further, if we let 𝑓 ( 𝑆 , ) ( 𝑂 , ) be an 𝑆 -fibration and ( 𝐸 , ) be an 𝑆 -subspace of ( 𝑂 , ) . Then 𝑆 -map 𝑓 | 𝑓 1 ( 𝐸 ) ( 𝑓 1 ( 𝐸 ) , ) ( 𝐸 , ) is an 𝑆 -fibration, where 𝑓 | 𝐸 will be the 𝑆 -fibration 𝑓 | 𝑓 1 ( 𝐸 ) . Thus, one by defining two 𝑆 -fibrations 𝑓 1 ( 𝑆 1 , 1 ) ( 𝑂 , ) and 𝑓 2 ( 𝑆 2 , 2 ) ( 𝑂 , ) , an 𝑆 -map ( 𝑆 1 , 1 ) ( 𝑆 2 , 2 ) and setting the relation 𝑓 2 = 𝑓 1 , then we can also study two 𝑆 -fiber maps ( 𝑆 1 , 1 ) ( 𝑆 2 , 2 ) and 𝑔 ( 𝑆 1 , 1 ) ( 𝑆 2 , 2 ) whether they are 𝑆 -fiber homotopic.

Theorem 1.1 (See [4]). Let 𝑋 , 𝑌 , and 𝑍 be topological spaces. Then the map 𝐹 𝑋 × 𝑌 𝑍 always gives rise to a map 𝐻 𝑋 𝐶 ( 𝑌 , 𝑍 ) by defining 𝐻 ( 𝑥 ) ( 𝑦 ) = 𝐹 ( 𝑥 , 𝑦 ) for all 𝑥 𝑋 , 𝑦 𝑌 . If 𝑌 is locally compact and regular space, then the map 𝐻 𝑋 𝐶 ( 𝑌 , 𝑍 ) always gives rise to map 𝐹 𝑋 × 𝑌 𝑍 by defining 𝐹 ( 𝑥 , 𝑦 ) = 𝐻 ( 𝑥 ) ( 𝑦 ) for all 𝑥 𝑋 , 𝑦 𝑌 .

By the natural topological semigroup, we mean a topological semigroup ( 𝑆 , 𝜋 𝑖 ) , where 𝜋 𝑖 is continuous associative multiplication on 𝑆 defined by 𝜋 1 ( 𝑥 , 𝑦 ) = 𝑥 and 𝜋 2 ( 𝑥 , 𝑦 ) = 𝑦 for all 𝑥 , 𝑦 𝑆 .

2. 𝑆 -Homotopy Extension Property

In this section, we extend the notions of an absolute retract (AR), an absolute neighborhood retract (ANR) and homotopy extension property in homotopy theory for topological spaces into their analogical structure in homotopy theory for topological semigroups. Also we show the relations among them.

Definition 2.1 (See [1]). Let ( 𝑂 , ) be an 𝑆 -subspace of topological semigroup ( 𝑆 , ) . The 𝑆 -retraction of ( 𝑆 , ) onto ( 𝑂 , ) is an 𝑆 -map 𝑟 ( 𝑆 , ) ( 𝑂 , ) such that 𝑟 ( 𝑠 ) = 𝑠 for all 𝑠 𝑂 . When there is an 𝑆 -retraction of ( 𝑆 , ) onto ( 𝑂 , ) , then we say that ( 𝑂 , ) is an 𝑆 -retract of ( 𝑆 , ) .

Definition 2.2. A topological semigroup ( 𝑆 , ) is called an 𝑆 -absolute retract ( 𝑆 -AR) for normal topological semigroups ( 𝑋 , ) in the class if for every closed 𝑆 -subspace ( 𝐴 , ) of ( 𝑋 , ) , any 𝑆 -map 𝑓 ( 𝐴 , ) ( 𝑆 , ) has an extension 𝑆 -map 𝐹 ( 𝑋 , ) ( 𝑆 , ) .

Definition 2.3. A topological semigroup ( 𝑆 , ) is called an 𝑆 -absolute neighborhood retract ( 𝑆 -ANR) for normal topological semigroups ( 𝑋 , ) in the class if for every closed 𝑆 -subspace ( 𝐴 , ) of ( 𝑋 , ) , any 𝑆 -map 𝑓 ( 𝐴 , ) ( 𝑆 , ) can be extended to an open neighborhood 𝑆 -subspace of ( 𝐴 , ) .

By using the above Definitions 2.2 and 2.3, we can easily state the following theorem.

Theorem 2.4. If ( 𝑆 , ) is an 𝑆 -AR (resp. 𝑆 -ANR), then 𝑆 is an AR (resp. ANR).

The converse of Theorem 2.4 above needs not to be true for example: If denotes the usual multiplication on the the unit closed interval 𝐼 . 𝐽 denotes the first half [ 0 , 1 / 2 ] of 𝐼 . It is clear that the topological semigroup ( 𝐽 , ) is a closed 𝑆 -subspace of a normal topological semigroup ( 𝐼 , ) . Also it is clear that 𝐽 is an ANR. Take 𝐴 = 𝐽 and 𝑋 = 𝐼 in Definition 2.3, let the identity 𝑆 -map 𝑖 ( 𝐽 , ) ( 𝐽 , ) has an extension 𝑆 -map 𝑅 ( 𝑀 , ) ( 𝐽 , ) , where ( 𝑀 , ) is an open neighborhood 𝑆 -subspace of ( 𝐽 , ) in ( 𝐼 , ) . In the relative usual topology on 𝐼 , we can get an element 𝑥 𝑀 and 𝑥 𝐽 , that is, 1 / 2 < 𝑥 1 . Since 𝑥 / 2 𝐽 and an 𝑆 -map 𝑅 would have to satisfy an impossible condition: 𝑥 2 𝑥 = 𝑟 2 1 = 𝑅 𝑥 2 1 = 𝑅 ( 𝑥 ) 𝑅 2 = 1 2 𝑅 ( 𝑥 ) ( 2 . 1 )

that implies 𝑅 ( 𝑥 ) = 𝑥 𝐽 . Hence ( 𝐽 , ) is not 𝑆 -ANR. Similarly, 𝐽 is an AR that but it is not 𝑆 -AR.

The converses of Theorem 2.4 are true for the collection of natural topological semigroups 𝒩 𝑖 as shown in the following theorem.

Theorem 2.5. A topological semigroup ( 𝑆 , 𝜋 𝑖 ) is an 𝑆 𝒩 𝑖 -AR (resp. an 𝑆 𝒩 𝑖 -ANR) for normal topological semigroups ( 𝑋 , 𝜋 𝑖 ) if and only if the topological space 𝑆 is AR (resp. ANR).

Proof. It is clear that the function 𝑓 ( 𝑋 , 𝜋 𝑖 ) ( 𝑆 , 𝜋 𝑖 ) is an 𝑆 -map if and only if the function 𝑓 𝑋 𝑆 is a continuous. Hence a topological semigroup ( 𝑆 , 𝜋 𝑖 ) is an 𝑆 𝒩 𝑖 -AR (resp. an 𝑆 𝒩 𝑖 -ANR) for normal topological semigroups ( 𝑋 , 𝜋 𝑖 ) if and only if the topological space 𝑆 is AR (resp. ANR).

Theorem 2.6. If ( 𝑆 , ) is an 𝑆 -AR, then 𝑃 ( 𝑆 , ) is also an 𝑆 -AR.

Proof. Let ( 𝐴 , ) be a closed 𝑆 -subspace of a normal topological semigroup ( 𝑋 , ) and let 𝐻 ( 𝐴 , ) 𝑃 ( 𝑆 , ) be an 𝑆 -map. Since 𝐼 is a locally compact and regular, then by Theorem 1.1, for 𝑡 𝐼 , we can define an 𝑆 -map 𝑓 𝑡 ( 𝐴 , ) ( 𝑆 , ) by 𝑓 𝑡 ( 𝑎 ) = 𝐻 ( 𝑎 ) ( 𝑡 ) , f o r 𝑎 𝐴 . ( 2 . 2 ) Since ( 𝑆 , ) is an 𝑆 -AR, then, for 𝑡 𝐼 , the 𝑆 -map 𝑓 𝑡 has an extension 𝑆 -map 𝐹 𝑡 ( 𝑋 , ) ( 𝑆 , ) . Hence 𝐻 has an extension 𝑆 -map 𝐹 ( 𝑋 , ) 𝑃 ( 𝑆 , ) defined by 𝐹 ( 𝑥 ) ( 𝑡 ) = 𝐹 𝑡 ( 𝑥 ) , f o r 𝑥 𝑋 , 𝑡 𝐼 . ( 2 . 3 ) Hence, 𝑃 ( 𝑆 , ) is an 𝑆 -AR.

Now in the following definitions, we give the concept of an 𝑆 -homotopy extension property in homotopy theory for topological semigroups.

Definition 2.7. Let ( 𝑆 , ) be a topological semigroup and ( 𝐴 , ) be an 𝑆 -subspace of ( 𝑆 , ) . We will mean by ( 𝐺 1 𝐴 , 𝐺 2 𝑆 , 𝑂 ) -maps for   ( 𝐴 , )   in ( 𝑆 , )   with respect to a topological semigroup ( 𝑂 , ) the two 𝑆 -maps 𝐺 1 𝐴 ( 𝐴 , ) 𝑃 ( 𝑂 , ) , 𝐺 2 𝑆 ( 𝑆 , ) ( 𝑂 , ) ( 2 . 4 ) such that 𝐺 1 𝐴 ( 𝑎 ) ( 0 ) = 𝐺 2 𝑆 ( 𝑎 ) f o r 𝑎 𝐴 . ( 2 . 5 )

Definition 2.8. Let ( 𝑆 , ) and ( 𝑂 , ) be two topological semigroups. A closed 𝑆 -subspace ( 𝐴 , ) of ( 𝑆 , ) is said to have 𝑆 -homotopy extension property in ( 𝑆 , )   with respect to ( 𝑂 , ) if any ( 𝐺 1 𝐴 , 𝐺 2 𝑆 , 𝑂 ) -maps can be extended to an 𝑆 -homotopy 𝐹 ( 𝑆 , ) 𝑃 ( 𝑂 , ) . That is, 𝐹 ( 𝑠 ) ( 0 ) = 𝐺 2 𝑆 ( 𝑠 ) , 𝐹 ( 𝑎 ) ( 𝑡 ) = 𝐺 1 𝐴 ( 𝑎 ) ( 𝑡 ) , f o r 𝑎 𝐴 , 𝑠 𝑆 , 𝑡 𝐼 . ( 2 . 6 )

Theorem 2.9. If a closed 𝑆 -subspace ( 𝐴 , ) of ( 𝑆 , ) has 𝑆 -homotopy extension property in ( 𝑆 , ) with respect to ( 𝑂 , ) , then 𝐴 has homotopy extension property in a space 𝑆 with respect to a space 𝑂 .

Proof. Let 𝐺 ( 𝑆 × { 0 } ) ( 𝐴 × 𝐼 ) 𝑂 be any continuous map. Define ( 𝐺 1 𝐴 , 𝐺 2 𝑆 , 𝑂 ) -maps for ( 𝐴 , 𝜋 1 ) in ( 𝑆 , 𝜋 1 ) with respect to a topological semigroup ( 𝑂 , 𝜋 1 ) by 𝐺 2 𝑆 ( 𝑠 ) = 𝐺 ( 𝑠 , 0 ) and 𝐺 1 𝐴 ( 𝑎 ) ( 𝑡 ) = 𝐺 ( 𝑎 , 𝑡 ) for all 𝑡 𝐼 ,  𝑠 𝑥 𝑆 ,  𝑎 𝐴 . Then by the hypothesis ( 𝐺 1 𝐴 , 𝐺 2 𝑆 , 𝑂 ) -maps can be extended to an 𝑆 -homotopy 𝐹 ( 𝑆 , 𝜋 1 ) 𝑃 ( 𝑂 , 𝜋 1 ) . That is, 𝐹 ( 𝑠 ) ( 0 ) = 𝐺 2 𝑆 ( 𝑠 ) , 𝐹 ( 𝑎 ) ( 𝑡 ) = 𝐺 1 𝐴 ( 𝑎 ) ( 𝑡 ) , f o r 𝑎 𝐴 , 𝑠 𝑆 , 𝑡 𝐼 . ( 2 . 7 ) Since 𝐼 is a locally compact and regular space, then by Theorem 1.1 the function 𝐻 𝑆 × 𝐼 𝑂 which is defined by 𝐻 ( 𝑠 , 𝑡 ) = 𝐹 ( 𝑠 ) ( 𝑡 ) for all 𝑡 𝐼 , 𝑠 𝑆 is continuous. Hence 𝐻 is an extension of 𝐺 . That is, 𝐴 has homotopy extension property in a space 𝑆 with respect to a space 𝑂 .

In the following theorem, we clarify that any closed 𝑆 -subspace ( 𝐴 , ) of a normal topological semigroup ( 𝑆 , ) has 𝑆 -homotopy extension property in ( 𝑆 , ) with respect to any 𝑆 -AR space ( 𝑂 , ) .

Theorem 2.10. Let ( O , ) be an 𝑆 -AR and ( 𝐴 , ) be a closed 𝑆 -subspace of a normal topological semigroup ( 𝑆 , ) . Then ( 𝐴 , ) has 𝑆 -homotopy extension property in ( 𝑆 , ) with respect to ( 𝑂 , ) .

Proof. Let that there is ( 𝐺 1 𝐴 , 𝐺 2 𝑆 , 𝑂 ) -maps. Since 𝐼 is a locally compact and regular, then by Theorem 1.1, for 𝑡 𝐼 { 0 } , define an 𝑆 -map 𝑓 𝑡 ( 𝐴 , ) ( 𝑂 , ) by 𝑓 𝑡 ( 𝑎 ) = 𝐺 1 𝐴 ( 𝑎 ) ( 𝑡 ) , f o r 𝑎 𝐴 . ( 2 . 8 ) Since ( 𝑂 , ) is an 𝑆 -AR, then for 𝑡 𝐼 { 0 } , the 𝑆 -map 𝑓 𝑡 has an extension 𝑆 -map 𝐹 𝑡 ( 𝑆 , ) ( 𝑂 , ) . While at the case 𝑡 = 0 , we can take 𝐹 0 = 𝐺 2 𝑆 . Hence there is an 𝑆 -map 𝐹 ( 𝑆 , ) 𝑃 ( 𝑂 , ) defined by 𝐺 𝐹 ( 𝑠 ) ( 𝑡 ) = 2 𝑆 𝐹 ( 𝑠 ) , f o r 𝑠 𝑆 , 𝑡 = 0 , 𝑡 ( 𝑠 ) , f o r 𝑠 𝑆 , 𝑡 𝐼 { 0 } . ( 2 . 9 ) Also we observe that 𝐹 ( 𝑎 ) ( 𝑡 ) = 𝐹 𝑡 ( 𝑎 ) = 𝑓 𝑡 ( 𝑎 ) = 𝐺 1 𝐴 ( 𝑎 ) ( 𝑡 ) , f o r 𝑎 𝐴 . ( 2 . 1 0 ) That is, ( 𝐴 , ) has 𝑆 -homotopy extension property in ( 𝑆 , ) with respect to ( 𝑂 , ) .

3. 𝑆 -Homotopy Extension Theorems

The general extension theorem for maps of any closed subsets of space 𝑋 into space 𝑌 discusses under which conditions on 𝑋 and 𝑌 , for every closed 𝐴 𝑋 , each map 𝑓 𝐴 𝑌 will be extendable over 𝑋 relative to 𝑌 . The concepts of an AR, an ANR, and homotopy extension property support this theorem in homotopy theory for topological spaces. Now in this section, we introduce the notions of homotopy extension theorem and fiber homotopy extension theorem in homotopy theory for topological semigroups via 𝑆 -AR property.

Lemma 3.1. Let ( 𝑆 , ) be a topological semigroup. For each path 𝛼 in ( 𝑆 , ) and 𝑟 𝐼 , let 𝛼 𝑟 be a path in ( 𝑆 , ) defined by 𝛼 𝑟 ( 𝑡 ) = 𝛼 [ ( 𝑟 , 𝑡 ) ] for all 𝑡 𝐼 , where 𝐼 × 𝐼 𝐼 be any map. Then the function 𝐻 𝑃 ( 𝑆 , ) 𝑃 ( 𝑃 𝑎 ( 𝑆 ) , 𝑝 ( ) ) defined by 𝐻 ( 𝛼 ) ( 𝑟 ) = 𝛼 𝑟 f o r 𝑟 𝐼 , 𝛼 𝑃 𝑎 ( 𝑆 ) ( 3 . 1 ) is an 𝑆 -map.

Proof. By Theorem 1.1 to prove that 𝐻 is continuous, it is sufficient to prove that the function 𝐹 𝑃 𝑎 ( 𝑆 ) × 𝐼 𝑃 𝑎 ( 𝑆 ) defined by 𝐹 ( 𝛼 , 𝑟 ) = 𝛼 𝑟 for all 𝑟 𝐼 , 𝛼 𝑃 𝑎 ( 𝑆 ) is continuous. Let ( 𝛽 , 𝑟 𝑜 ) 𝑃 𝑎 ( 𝑆 ) × 𝐼 and 𝑊 ( 𝐾 , 𝑈 ) be a neighborhood of 𝛽 𝑟 𝑜 in 𝑃 𝑎 ( 𝑆 ) . Then 𝛽 ( ( { 𝑟 𝑜 } × 𝐾 ) ) = 𝛽 𝑟 𝑜 ( 𝐾 ) 𝑈 . By the continuity of 𝛽 and 𝑈 is an open set containing 𝛽 ( ( { 𝑟 𝑜 } × 𝐾 ) ) , there is open set 𝐺 in 𝐼 such that ( { 𝑟 𝑜 } × 𝐾 ) 𝐺 and 𝛽 ( 𝐺 ) 𝑈 . Also by the continuity of , there are two open sets 𝐼 𝑟 𝑜 and 𝐼 𝐾 in 𝐼 such that 𝑟 𝑜 × 𝐾 𝐼 𝑟 𝑜 × 𝐼 𝐾 𝐼 , 𝑟 𝑜 × 𝐼 𝐾 𝐺 . ( 3 . 2 ) Since 𝐼 𝑟 𝑜 is an open set in 𝐼 containing 𝑟 𝑜 , then there is a positive number 𝜖 > 0 such that 𝑟 𝑜 𝑟 𝑜 𝜖 3 , 𝑟 𝑜 + 𝜖 3 𝑟 𝐷 = 𝑜 𝜖 2 , 𝑟 𝑜 + 𝜖 2 𝑟 𝑜 𝜖 , 𝑟 𝑜 + 𝜖 𝐼 𝑟 𝑜 . ( 3 . 3 ) Since 𝐷 and 𝐾 are compact sets in 𝐼 and is a continuous, then ( 𝐷 × 𝐾 ) is also a compact set in 𝐼 . Now consider that 𝑊 ( ( 𝐷 × 𝐾 ) , 𝑈 ) × ( 𝑟 𝑜 𝜖 / 3 , 𝑟 𝑜 + 𝜖 / 3 ) a neighborhood of ( 𝛽 , 𝑟 𝑜 ) in 𝑃 𝑎 ( 𝑆 ) × 𝐼 . Hence for 𝑟 ( 𝛼 , 𝑟 ) 𝑊 ( ( 𝐷 × 𝐾 ) , 𝑈 ) × 𝑜 𝜖 3 , 𝑟 𝑜 + 𝜖 3 , 𝐹 ( 𝛼 , 𝑟 ) ( 𝐾 ) = 𝛼 𝑟 [ ] ( 𝐾 ) = 𝛼 ( ( { 𝑟 } × 𝐾 ) ) 𝛼 ( 𝐷 × 𝐾 ) 𝑈 . ( 3 . 4 ) That is, 𝐹 ( 𝛼 , 𝑟 ) 𝑊 ( 𝐾 , 𝑈 ) . Hence the function 𝐹 is continuous.
Now for 𝛼 , 𝛽 𝑃 ( 𝑆 , ) , [ ] [ ] { 𝐻 𝑝 ( ) ( 𝛼 , 𝛽 ) ( 𝑟 ) } ( 𝑡 ) = 𝑝 ( ) ( 𝛼 , 𝛽 ) 𝑟 = [ 𝑝 ] [ ] 𝛼 ( 𝑡 ) ( ) ( 𝛼 , 𝛽 ) ( ( 𝑟 , 𝑡 ) ) = 𝛼 ( ( 𝑟 , 𝑡 ) ) , 𝛽 ( ( 𝑟 , 𝑡 ) ) = 𝑟 ( 𝑡 ) , 𝛽 𝑟 ( = 𝑝 𝛼 𝑡 ) ( ) 𝑟 , 𝛽 𝑟 [ ] ( 𝑡 ) = { 𝑝 ( 𝑝 ( ) ) 𝐻 ( 𝛼 ) , 𝐻 ( 𝛽 ) ( 𝑟 ) } ( 𝑡 ) . ( 3 . 5 ) Hence 𝐻 [ 𝑝 ( ) ( 𝛼 , 𝛽 ) ] = 𝑝 ( 𝑝 ( ) ) [ 𝐻 ( 𝛼 ) , 𝐻 ( 𝛽 ) ] , that is, 𝐻 is an 𝑆 -map.

The following theorem clarifies the notion of 𝑆 -fiber homotopy extension theorem in homotopy theory for topological semigroups via 𝑆 -AR property

Theorem 3.2. Let 𝑓 ( 𝑆 , ) ( 𝑂 , ) be an 𝑆 -regular fibration with an 𝑆 -AR ( 𝑆 , ) . Let ( 𝐴 , ) be a closed 𝑆 -subspace of a normal topological semigroup ( 𝑋 , ) . If there is ( 𝐺 1 𝐴 , 𝐺 2 𝑋 , 𝑆 ) -maps such that 𝑓 𝐺 1 𝐴 𝐺 ( 𝑎 ) ( 𝑡 ) = 𝑓 1 𝐴 ( 𝑎 ) ( 0 ) f o r 𝑎 𝐴 , 𝑡 𝐼 , ( 3 . 6 ) then there is an 𝑆 -map 𝐻 ( 𝑋 , ) 𝑃 ( 𝑆 , ) such that 𝐻 is an extension of 𝐺 1 𝐴 , 𝐻 0 = 𝐺 2 𝑋 and 𝑓 [ ] [ ] 𝐻 ( 𝑥 ) ( 𝑡 ) = 𝑓 𝐻 ( 𝑥 ) ( 0 ) f o r 𝑥 𝑋 , 𝑡 𝐼 . ( 3 . 7 )

Proof. Since ( 𝑆 , ) is an 𝑆 -AR and ( 𝐴 , ) is a closed 𝑆 -subspace of a normal topological semigroup ( 𝑋 , ) , then, by Theorem 2.10, the 𝑆 -map 𝐺 1 𝐴 can be extended to 𝑆 -map 𝐺 ( 𝑋 , ) 𝑃 ( 𝑆 , ) such that 𝐺 0 = 𝐺 2 𝑋 . Now for 𝛼 𝑃 ( 𝑆 , ) and 𝑟 𝐼 , we can define the path 𝛼 𝑟 in 𝑃 ( 𝑆 , ) by 𝛼 𝑟 ( 𝑡 ) = 𝛼 [ ( 1 𝑡 ) 𝑟 ] for all 𝑡 𝐼 . Then by Lemma 3.1, the function 𝑃 ( 𝑆 , ) 𝑃 ( 𝑃 𝑎 ( 𝑆 ) , 𝑝 ( ) ) ( 𝛼 ) ( 𝑟 ) 𝛼 𝑟 ( 3 . 8 ) is an 𝑆 -map. Hence, we can define the 𝑆 -map 𝐻 ( 𝑋 , ) 𝑃 ( 𝑆 , ) by 𝐻 ( 𝑥 ) ( 𝑡 ) = 𝐿 𝑓 𝐺 ( 𝑥 ) ( 𝑡 ) , 𝑓 𝐺 ( 𝑥 ) 𝑡 ( 1 ) f o r 𝑥 𝑋 , 𝑡 𝐼 . ( 3 . 9 ) Firstly, we show that 𝐻 0 = 𝐺 2 𝑋 . By the 𝑆 -regularity of 𝐿 𝑓 and since 𝐺 is an extension for 𝐺 1 𝐴 , we observe that for 𝑥 𝑋 , 𝐻 ( 𝑥 ) ( 0 ) = 𝐿 𝑓 𝐺 ( 𝑥 ) ( 0 ) , 𝑓 𝐺 ( 𝑥 ) 0 ( 1 ) = 𝐿 𝑓 ( 𝐺 ( 𝑥 ) ( 0 ) , 𝑓 𝐺 ( 𝑥 ) ( 0 ) ) ( 1 ) = 𝐿 𝑓 = 𝐺 ( 𝑥 ) ( 0 ) , 𝑓 𝐺 ( 𝑥 ) ( 0 ) ( 1 ) 𝐺 ( 𝑥 ) ( 0 ) ( 1 ) = 𝐺 ( 𝑥 ) ( 0 ) = 𝐺 2 𝑋 ( 𝑥 ) . ( 3 . 1 0 ) Secondly, we show that 𝐻 is an extension of 𝐺 1 𝐴 . Since 𝐺 is an extension for 𝐺 1 𝐴 and by the hypothesis, we get that 𝑓 𝐺 ( 𝑎 ) 𝑟 [ 𝐺 ] 𝐺 ( 𝑡 ) = 𝑓 ( 𝑎 ) ( ( 1 𝑡 ) 𝑟 ) = 𝑓 1 𝐴 𝐺 ( 𝑎 ) ( ( 1 𝑡 ) 𝑟 ) = 𝑓 1 𝐴 𝐺 ( 𝑎 ) ( 0 ) = 𝑓 1 𝐴 [ ] = ( 𝑎 ) ( 𝑟 ) = 𝑓 𝐺 ( 𝑎 ) ( 𝑟 ) 𝑓 𝐺 ( 𝑎 ) ( 𝑟 ) ( 𝑡 ) , ( 3 . 1 1 ) for all 𝑎 𝐴 and 𝑟 , 𝑡 𝐼 . Hence, by the 𝑆 -regularity of 𝐿 𝑓 , we get that 𝐻 ( 𝑎 ) ( 𝑟 ) = 𝐺 ( 𝑎 ) ( 𝑟 ) = 𝐺 1 𝐴 ( 𝑎 ) ( 𝑟 ) f o r 𝑟 𝐼 , 𝑎 𝐴 . ( 3 . 1 2 ) Hence 𝐻 is an extension for 𝐺 1 𝐴 . Finally, we also observe that 𝑓 [ 𝐻 ] 𝐿 ( 𝑥 ) ( 𝑡 ) = 𝑓 𝑓 𝐺 ( 𝑥 ) ( 𝑡 ) , 𝑓 𝐺 ( 𝑥 ) 𝑡 = ( 1 ) 𝑓 𝐺 ( 𝑥 ) 𝑡 ( 1 ) = 𝑓 𝐺 ( 𝑥 ) 𝑡 [ ] 𝐺 ( 1 ) = 𝑓 𝐺 ( 𝑥 ) ( 0 ) = 𝑓 2 𝑋 ( [ ] , 𝑥 ) = 𝑓 𝐻 ( 𝑥 ) ( 0 ) ( 3 . 1 3 ) for all 𝑥 𝑋 , 𝑡 𝐼 .

Now we can give the other rephrasing of above theorem in the following corollary.

Corollary 3.3. Let 𝑓 ( 𝑆 , ) ( 𝑂 , ) be an 𝑆 -regular fibration with an 𝑆 -AR ( 𝑆 , ) and ( 𝐴 , ) be a closed 𝑆 -subspace of a normal topological semigroup ( 𝑋 , ) . Let 𝑘 1 , 𝑘 2 ( 𝐴 , ) ( 𝑆 , ) be two 𝑆 -maps and 𝑅 ( 𝐴 , ) 𝑃 ( 𝑆 , ) be an 𝑆 -homotopy between them such that 𝑓 [ ] [ ] 𝑅 ( 𝑎 ) ( 𝑡 ) = 𝑓 𝑅 ( 𝑎 ) ( 0 ) f o r 𝑎 𝐴 , 𝑡 𝐼 . ( 3 . 1 4 ) If 𝑘 1 has an extension 𝑆 -map 𝐾 1 to all of ( 𝑋 , ) , then 𝑘 2 has an extension 𝑆 -map 𝐾 2 to all of ( 𝑋 , ) . Also there is an 𝑆 -homotopy 𝐻 ( 𝑋 , ) 𝑃 ( 𝑆 , ) between 𝐾 1 and 𝐾 2 such that 𝐻 is an extension of 𝑅 and 𝑓 [ ] [ ] 𝐻 ( 𝑥 ) ( 𝑡 ) = 𝑓 𝐻 ( 𝑥 ) ( 0 ) f o r 𝑥 𝑋 , 𝑡 𝐼 . ( 3 . 1 5 )

Proof. Since 𝑘 1 has an extension 𝑆 -map 𝐾 1 to all of ( 𝑋 , ) , that is, 𝐾 1 ( 𝑎 ) = 𝑘 1 ( 𝑎 ) = 𝑅 ( 𝑎 ) ( 0 ) f o r 𝑎 𝐴 . ( 3 . 1 6 ) Hence there is ( 𝑅 , 𝐾 1 , 𝑆 ) -maps with the property 𝑓 [ ] [ ] 𝑅 ( 𝑎 ) ( 𝑡 ) = 𝑓 𝑅 ( 𝑎 ) ( 0 ) f o r 𝑎 𝐴 , 𝑡 𝐼 . ( 3 . 1 7 ) Then by Theorem 3.2, there is an 𝑆 -map 𝐻 ( 𝑋 , ) 𝑃 ( 𝑆 , ) such that 𝐻 is an extension of 𝑅 , 𝐻 0 = 𝐾 1 and 𝑓 [ ] [ ] 𝐻 ( 𝑥 ) ( 𝑡 ) = 𝑓 𝐻 ( 𝑥 ) ( 0 ) f o r 𝑥 𝑋 , 𝑡 𝐼 . ( 3 . 1 8 ) Also we can define the 𝑆 -map 𝐾 2 ( 𝑋 , ) ( 𝑆 , ) by 𝐾 2 ( 𝑥 ) = 𝐻 ( 𝑥 ) ( 1 ) for all 𝑥 𝑋 . Then 𝐻 is an 𝑆 -homotopy between 𝐾 1 and 𝐾 2 and also 𝐾 2 ( 𝑎 ) = 𝐻 ( 𝑎 ) ( 1 ) = 𝑘 2 ( 𝑎 ) f o r 𝑎 𝐴 . ( 3 . 1 9 ) That is, 𝐾 2 is an extension of 𝑘 2 .

The following theorem clarifies the notion of 𝑆 -fiber homotopy extension theorem via 𝑆 -AR property.

Theorem 3.4. Let 𝑓 1 ( 𝑆 1 , 1 ) ( 𝑂 , ) and 𝑓 2 ( 𝑆 2 , 2 ) ( 𝑂 , ) be two 𝑆 -fibrations with an 𝑆 -AR ( 𝑆 2 , 2 ) , a normal ( 𝑆 1 , 1 ) and ( 𝐸 , ) be a closed 𝑆 -subspace of ( 𝑂 , ) . Let 𝑘 , 𝑘 ( 𝑓 1 1 ( 𝐸 ) , 1 ) ( 𝑓 2 1 ( 𝐸 ) , 2 ) be two 𝑆 -fiber maps between two 𝑆 -fibrations 𝑓 1 | 𝐸 and 𝑓 2 | 𝐸 such that 𝐾 𝑓 𝐾 . If 𝑘 is extendable to a full 𝑆 -fiber map 𝐾 ( 𝑆 1 , 1 ) ( 𝑆 2 , 2 ) , then 𝑘 is extendable to a full 𝑆 -fiber map 𝐾 ( 𝑆 1 , 1 ) ( 𝑆 2 , 2 ) and 𝐾 𝑓 𝐾 .

Proof. Since 𝐾 𝑓 𝐾 , then there is an 𝑆 homotopy 𝑓 𝑅 1 1 ( 𝐸 ) , 1 𝑓 𝑃 2 1 ( 𝐸 ) , 2 𝑆 𝑃 2 , 2 , ( 3 . 2 0 ) such that 𝑅 ( 𝑠 ) ( 0 ) = 𝑘 ( 𝑠 ) , 𝑅 ( 𝑠 ) ( 1 ) = 𝑘 ( 𝑠 ) and 𝑓 2 [ 𝑅 ] ( 𝑠 ) ( 𝑡 ) = 𝑓 1 ( 𝑠 ) = 𝑓 2 [ 𝑘 ] ( 𝑠 ) = 𝑓 2 [ 𝑅 ] ( 𝑠 ) ( 0 ) f o r 𝑠 𝑓 1 1 ( 𝐸 ) , 𝑡 𝐼 . ( 3 . 2 1 ) Since 𝑘 is extendable to a full 𝑆 -fiber map 𝐾 , then there is ( 𝐾 , 𝑅 , 𝑆 2 ) -maps. Since ( 𝑓 1 1 ( 𝐸 ) , 1 ) is a closed 𝑆 -subspace of a normal ( 𝑆 1 , 1 ) , we can apply Theorem 3.2 on 𝑆 -fibration 𝑓 2 , where 𝐴 = 𝑓 1 1 ( 𝐸 ) and 𝑋 = ( 𝑆 1 , 1 ) . Hence there is an 𝑆 -map 𝐻 ( 𝑆 1 , 1 ) 𝑃 ( 𝑆 2 , 2 ) such that 𝐻 is an extension of 𝑅 , 𝐻 0 = 𝐾 and 𝑓 2 [ ] 𝐻 ( 𝑠 ) ( 𝑡 ) = 𝑓 2 [ ] 𝐻 ( 𝑠 ) ( 0 ) f o r 𝑠 𝑆 1 , 𝑡 𝐼 . ( 3 . 2 2 ) Now we can define 𝐾 by 𝐾 = 𝐻 1 . Hence 𝐻 is an 𝑆 -homotopy between 𝐾 and 𝐾 . From (3.22) and since 𝐾 is an 𝑆 -fiber map, we get that 𝐾 is an 𝑆 -fiber map and 𝑓 2 [ ] 𝐻 ( 𝑠 ) ( 𝑡 ) = 𝑓 2 [ ] 𝐻 ( 𝑠 ) ( 0 ) = 𝑓 2 [ ] 𝐾 ( 𝑠 ) = 𝑓 1 ( 𝑠 ) f o r 𝑠 𝑆 1 , 𝑡 𝐼 . ( 3 . 2 3 ) Hence 𝐾 𝑓 𝐾 .

References

  1. Z. Čerin, “Homotopy theory for topological semigroups,” Topology and Its Applications, vol. 123, no. 1, pp. 57–68, 2002. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet
  2. M. V. Lawson, J. Matthews, and T. Porter, “The homotopy theory of inverse semigroups,” International Journal of Algebra and Computation, vol. 12, no. 6, pp. 755–790, 2002. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet
  3. G. González, “Locally generated semigroups,” Divulgaciones Matemáticas, vol. 4, no. 1-2, pp. 5–20, 1996. View at Zentralblatt MATH
  4. A. Hatcher, Algebraic Topology, Cambridge University Press, Cambridge, UK, 2002.