LINEAR COHYPERSUBSTITUTIONS AND GENERALIZED LINEAR COHYPERSUBSTITUTIONS OF TYPE τ = (n)
JULALUK BOONSOL
A Thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Mathematics
at Mahasarakham University October 2019
All rights reserved by Mahasarakham University
LINEAR COHYPERSUBSTITUTIONS AND GENERALIZED LINEAR COHYPERSUBSTITUTIONS OF TYPE τ = (n)
JULALUK BOONSOL
A Thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Mathematics
at Mahasarakham University October 2019
All rights reserved by Mahasarakham University
The examining committee has unanimously approved this thesis, submitted by Miss Julaluk Boonsol, as a partial fulfillment of the requirements for the Master of Science in Mathematics at Mahasarakham University.
Examining Committee
. . . Chairman
(Asst. Prof. Somsak Lekkoksung, Dr.rer.nat.) (External expert)
. . . Committee (Asst. Prof. Kittisak Saengsura, Dr.rer.nat.) (Advisor)
. . . Committee
(Asst. Prof. Ananya Anantayasethi, Dr.rer.nat.) (Faculty graduate committee)
. . . Committee
(Asst. Prof. Jeeranunt Khampakdee, Ph.D.) (Faculty graduate committee)
Mahasarakham University has granted approval to accept this thesis as a partial fulfillment of the requirements for the Master of Science in Mathematics.
. . . . (Prof. Pairot Pramual, Ph.D.) (Asst. Prof. Krit Chaimoon, Ph.D.) Dean of the Faculty of Science Dean of Graduate School
. . . 2019
ACKNOWLEDGEMENTS
It would be difficult to name all to whom I owe thanks for their assistance and help with this work. However special acknowledgement is extended to the following:
I gratefully acknowledge the financial support of the Science Achievement Scholarship of Thailand (SAST).
I express my gratitude to my advisor, Asst. Prof. Kittisak Saengsura, who introduced me to research, for his help and constant encouragement throughout the course of this research. I am the most grateful for his teaching and advice. This thesis would not have been completed without all the support that I have always received from him.
I would like to thanks Asst. Prof. Somsak Lekkoksung, Rajamangala University of Technology Isan Khonkaen Campus Khonkaen Thailand, Asst. Prof. Ananya Anantayasethi, and Asst. Prof. Jeeranunt Khampakdee, Department of Mathematics, Faculty of Science, Mahasarakham University for their kind comments and support as members of dissertation committee.
I express my family for their understanding and support me for everything until this study completion.
Finally, thanks are due to all of the staffs in the Science Achievement Scholarship of Thailand (SAST) and Department of Mathematics, Faculty of Science, Mahasarakham University, for their warm welcome.
Julaluk Boonsol
i
ii
TITLE Linear Cohypersubstitutions and Generalized Linear Cohypersubstitutions of type τ = (n)
CANDIDATE Miss Julaluk Boonsol
DEGREE Master of Science MAJOR Mathematics ADVISOR Asst. Prof. Kittisak Saengsura, Dr.rer.nat.
UNIVERSITY Mahasarakham University YEAR 2019
ABSTRACT
A coterm of type τ was introduced by K. Denecke and K. Saengsura in 2009 as a compound of two disjoint sets {fi |i∈I} and {eni |n ∈N,0≤ i≤n−1}. A linear coterm t is a coterm which the element of {eni | n ∈ N,0 ≤ i ≤ n−1} occurring in t are all different. In this research, we study some semigroup properties of linear cohypersubstitutions and generalized linear cohypersubstitutions of type τ = (n), for instance: idempotents, regulars and Green’s relations.
Keywords: linear cohypersubstitution, generalized linear cohypersubstitution, idempotent, regular, Green’s relations.
CONTENTS
Page
Acknowledgements i
Abstract ii
Contents iii
Chapter 1 Introduction 1
1.1 Background 1
Chapter 2 Preliminaries 4
2.1 Semigroups 4
2.2 Green’s relations 5
2.3 Algebras 6
2.4 Coalgebras 9
Chapter 3 Linear cohypersubstitutions of type τ = (n) 13 3.1 Monoid of linear cohypersubstitutions of type τ = (n) 13 3.2 Idempotent and some regular elements of linear cohypersubstitutions
of type τ = (n) 17
3.3 Green’s relations of linear cohypersubstitutions of type
τ = (n) 22
Chapter 4 Generalized linear cohypersubstitutions of type τ = (n) 29 4.1 Monoid of generalized linear cohypersubstitutions of
type τ = (n) 29
4.2 Idempotent and some regular elements of generalized linear cohypersubstitutions of type τ = (n) 41 4.3 Green’s relations of generalized linear cohypersubstitutions
of type τ = (n) 45
Chapter 5 Conclusions 51
References 58
Biography 61
iii
CHAPTER 1 INTRODUCTION
1.1 Background
In this dissertation, we mainly pay attention to the concept of cohypersubstitutions and generalized cohypersubstitutions.
Let A be a non-empty set and n be a positive integer. The n-th copower Atn of A is the union of n disjoint copies of A; formally, we define Atn as the cartesian product Atn :=n×A, where n :={0, ..., n−1}. An element (i, a) in this copower corresponds to the element a in the i-th copy of A, for 1≤i≤n. A co-operation on A is a mapping fA : A → Atn for some n ≥ 1; the natural number n is called the arity of the co-operation fA. We also need to recall that any n-ary co-operation fA on set A can be uniquely expressed as a pair (f1A, f2A) of mappings, f1A :A→ n and f2A : A → A; the first mapping gives the labelling used by fA in mapping elements to copies of A, and the second mapping tells us what element of A any element is mapped to.
We shall denote by cO(n)A = {fA|A → Atn} the set of all n-ary co-operations defined on A, and by cOA := ∪n≥1cOA(n) the set of all finitary co-operations defined on A. An indexed coalgebra is a pair(A; (fiA)i∈I), wherefiA is an ni-ary co-operation defined onA,andτ = (ni)i∈I forni ≥1is called the type of the coalgebra. Coalgebras were studied by Drbohlav [10]. In [1], the following superposition of co-operations was introduced. If fA ∈ cO(n)A and gA1, ..., gnA ∈ cOA(k), then the k-ary co-operation fA[g1A, ..., gAn] :A→Atk is defined by
a7→((gAfA
1(a))1(f2A(a)),(gfAA
1(a))2(f2A(a))),
for all a ∈ A. The co-operation fA[gA1, ..., gnA] is called the superposition of fA and g1A, ..., gnA. It will also be denoted by compnk(fA, g1A, ..., gAn).
1
2 The injection co-operations ιn,Ai :A → Atn are special co-operations which are defined for each 0≤i≤n−1 by ιn,Ai :A→Atn with a 7→(i, a) for all a ∈A. Then we get a many-sorted algebra
((cOA(n))n≥1,(compnk)k,n≥1,(ιn,Ai )0≤i≤n−1),
called the clone of co-operations on A. In [1], it is mentioned that this algebra is a clone, i.e., it satisfies the three clone axioms (C1),(C2),(C3). In [8], K. Denecke and K. Saengsura gave a full proof of this fact.
In 2016, D. Boonchari and K. Saengsura studied the monoid of cohypersubstitutions of type τ = (n) and characterized all idempotent and regular elements of set of all cohypersubstitutions (see [2]). The concept of cohypersubstitutions can be extended into the concept of generalized cohypersubstitutions defined by S. Jermjitpornchai and N. Saengsura in 2013 which we could find an original literature in [12]. Similar to the original concept of cohypersubstitutions, the set of all generalized cohypersubstitutions together with a special binary operation forms a monoid.
For our purposes, we introduce the notions of linear cohypersubstitutions of type τ = (n), then idempotent, some regular elements and Green’s relations of linear cohypersubstitutions of type τ = (n) are studied. Moreover, we study some properties of generalized linear cohypersubstitutions of type τ = (n). We devide our thesis into 5 chapters, as follows:
In the first chapter, the introduction was presented.
In chapter 2, we present some basic concepts and results of algebras and coalgebras without proof which are needed in the subsequent chapters.
In chapter 3, we introduce the concept of linear cohypersubstitutions of type τ = (n). We also study the properties of idempotent, regular elements and Green’s relations of monoid of linear cohypersubstitutions of type τ = (n).
3
In chapter 4, we introduce the concept of generalized linear cohypersubstitutions of type τ = (n). We also study the properties of idempotent, regular elements and Green’s relations of monoid of generalized linear cohypersubstitutions of typeτ = (n).
In the last chapter, we summarize results of our study.
CHAPTER 2 PRELIMINARIES
In this chapter, we will give some definitions, notations, dealing with some preliminaries and some useful results that will be duplicated in later chapters.
2.1 Semigroups
Definition 2.1.1. [4] Let S be a non-empty set, and ∗ said to be a binary operation on S, if a∗b is defined for all a, b∈S.
Definition 2.1.2. [4] A groupoid (S,∗) is defined as a non-empty set S on which a binary operation ∗, by which we mean a map ∗:S×S → S is defined. We say that (S,∗) is a semigroup if the operation ∗ is associative, that is to say, if for all x, y and z in S,
((x∗y)∗z) = (x∗(y∗z)).
Definition 2.1.3. [4] If a semigroup (S,∗) contains an element 1 with the property that, for all x in S,
x∗1 = 1∗x=x,
we say that 1 is an identity element (or just an identity) of S, and that (S,∗) is a semigroup with identity or (more usually) a monoid.
Definition 2.1.4. [4] An element a of a semigroup (S,∗) is called idempotent if a∗a=a.
Definition 2.1.5. [4] An element a of a semigroup (S,∗) is called regular if there exists x in S such that a∗x∗a =a. The semigroup (S,∗) is called regular if all its elements are regular.
4
5 2.2 Green’s relations
Definition 2.2.1. [4] For semigroup S, we define
S1 :=
S if S has an identity element S∪ {1} otherwise.
Definition 2.2.2. [4] If a is an element of a semigroup S without identity then Sa need not contain a. The following notations will be standard:
S1a =Sa∪ {a}
aS1 =aS∪ {a}
S1aS1 =SaS∪Sa∪aS ∪ {a}.
Notice that S1a, aS1 and S1aS1 are all subsets of S-they do not contain the element 1.
Remark. We call S1a, aS1 and S1aS1 the principal left ideal generated by a.
Definition 2.2.3. [4] An equivalence L on S is defined by the rule that aLb if and only ifa andb generate the same principal left ideal, that is, if and only if S1a =S1b.
Similarly, we define the equivalence R by the rule that aRb if and only if aS1 =bS1. Proposition 2.2.4. [4] Let a, b be elements of a semigroup S. Then aLb if and only if there exist x, y in S1 such that xa=b, yb=a. Also, aRb if and only if there exist u, v in S1 such that au=b, bv =a.
Definition 2.2.5. [4] Let a, b be elements of a semigroup S. An equivalence H on S is defined by the rule that aHb if and only if aLb and aRb.
Definition 2.2.6. [4] Let a, b be elements of a semigroup S. An equivalence D on S is defined by the rule that aDb if and only if there exists c in S such that aLc and cRb.
Definition 2.2.7. [4] An equivalence J on S is defined by the rule that aJb if and only if S1aS1 =S1bS1, that is to say, if and only if there exist x, y, u, v in S1 such that xay =b, ubv =a.
6 2.3 Algebras
Denote the set of all positive integers by N+. That is, N+:={1,2,3, ...}.
Definition 2.3.1. [9] An algebra is a pair A := (A;FA) consisting of a non-empty set A with the set FA = {fiA|i ∈ I} of operations defined on A indexed by some non-empty set I, where for each i∈I, fiA is an ni-ary operation defined on A.
Let A:= (A;FA) be an algebra. The set A is called the universe of the algebra, and the fiA are the basic operations where i ∈ I. Moreover, a non-negative integer ni is called the arity of the basic operation fiA, and the sequence of all arities τ := (ni)i∈I is called the similarity type, for short type, of the algebra A. However, we may consider FA as (fiA)i∈I for some index set I. For convenience, we may write (A; (fiA)i∈I) instead of(A;FA). Denoted byAlg(τ)the class of all algebras of a given type τ.
Let Xn :={x1, ..., xn}, where n∈N+. The element xi in Xn is called variable.
Definition 2.3.2. [9] An n-ary term of type τ is defined in the following inductive ways.
(1) Every variable in Xn is an n-ary term.
(2) If t1, ..., tni are n-ary terms and fi is ni-ary operation symbol, then fi(t1, ..., tni) is an n-ary term.
We denote the set of all n-ary terms of type τ by Wτ(Xn). The set Wτ(Xn) is the smallest set containing Xn and is closed under finite application of (2). Denoted by Wτ(X) the set of all terms of type τ. That is,
Wτ(X) := [
n∈N+
Wτ(Xn).
Definition 2.3.3. [3] Let m, n∈ N+. The operation Snm :Wτ(Xm)×(Wτ(Xn))m → Wτ(Xn) is defined inductively as follows:
(1) Snm(xi, t1, ..., tm) := ti, where 1≤i≤m.
7 (2) Snm(fi(s1, ..., sni), t1, ..., tm) := fi(Snm(s1, t1, ..., tm), ..., Snm(sni, t1, ..., tm)) and
assume that Snm(sj, t1, ..., tm) is already defined for all 1≤j ≤ni.
Definition 2.3.4. [6] A hypersubstitution is a mapping from the set of all operation symbols to the set of terms preserving arity. Denoted by Hyp(τ) the set of all hypersubstitutions of type τ.
Any hypersubstitution σ can be extened to mapping σˆ :Wτ(X)→Wτ(X) defined by the following steps:
(1) σ[x] :=ˆ x for all variable x∈X.
(2) σ[fˆ i(t1, ..., tni)] := Snni(σ(fi),σ[tˆ 1], ...,σ[tˆ ni]) if fi(t1, ..., tni) ∈ Wτ(Xn) and ˆ
σ[tj]∈Wτ(Xn) is already defined for all 1≤j ≤ni.
In [6], the author showed that the set Hyp(τ) of all hypersubstitutions of type τ endowed with a binary operation ◦h, defined by σ1◦hσ2 := ˆσ1◦σ2, forms a monoid, where σid, defined by σid(fi) := fi(x1, ..., xni), is its neutral element. The algebraic system Hyp(τ) := (Hyp(τ);◦h, σid) is a monoid.
Definition 2.3.5. [11] Let n ∈ N. An (n+ 1)-ary generalized superposition is an (n+ 1)-ary operation Sn on the set Wτ(X) defined by the following steps:
(1) if t =xi ∈Xn, then Sn(t, t1, ..., tn) :=ti, (2) if t =xj ∈X\Xn, then Sn(t, t1, ..., tn) := xj,
(3) if t = fi(s1, ..., sni) and assume that Sn(sj, t1, ..., tn) is already defined for all 1≤j ≤n, then
Sn(t, t1, ..., tn) :=fi(Sn(s1, t1, ..., tn), ..., Sn(sni, t1, ..., tn)).
Then we define the extension σˆ :Wτ(X)→Wτ(X) by the following steps:
(1) If t ∈X, then σ[t] :=ˆ t.
(2) Ift=fi(t1, ..., tni),thenσ[t] :=ˆ Sni(σ(fi),σ[tˆ 1], ...,σ[tˆ ni]),whereσ[tˆ j]is already defined for all 1≤j ≤ni.
8 This extension is uniquely determined and allows us to define a multiplication, denoted by ◦G, on the set HypG(τ) of all generalized hypersubstitions of type τ by
σ1◦Gσ2 := ˆσ1◦σ2
where ◦ denotes the usual composition of mappings. Let σid be the identity hypersubstitution which maps eachni-ary operation symbolfi to the termfi(x1, ..., xni).
The algebraic properties HypG(τ) := (HypG(τ);◦G, σid) is a monoid.
Let var(t) be the set of variables occurring in t.
Definition 2.3.6. [5] Ann-ary linear term of typeτ is defined by induction as follows:
(i) xj ∈Xn is an n-ary linear term of type τ for all j ∈ {1, ..., n}.
(ii) If t1, ..., tni are n-ary linear terms of type τ and var(tj)∩var(tk) = ∅ for all 1≤j < k ≤ni, then fi(t1, ..., tni) is an n-ary linear term of type τ.
(iii) The set Wτlin(Xn) of all n-ary linear terms of type τ is the least set containing x1, ..., xn and closed under finite applications of (ii).
Wτlin(X) denotes the union of all Wτlin(Xn) for n∈N+.
Definition 2.3.7. [5] The partial many-sorted mappingSmn :Wτlin(Xm)×(Wτlin(Xn))m (→Wτlin(Xn) by
Smn(t, s1, ..., sm) :=
Snm(t, s1, ..., sm) if var(sj)∩var(sk) = ∅ for all 1≤j < k≤m, not defined otherwise
for m, n∈N+.
Definition 2.3.8. [5] A linear hypersubstitution is a mapping from the set of all operation symbols to the set of linear terms preserving arity. Denoted by Hyplin(τ) the set of all linear hypersubstitutions of type τ.
Any linear hypersubstitutionαcan be extened to mapping αˆ:Wτlin(X)→Wτlin(X) defined by the following steps:
(1) α[x] :=ˆ x for all variable x∈X.
9 (2) α[fˆ i(t1, ..., tni)] := Snni(α(fi),α[tˆ 1], ...,α[tˆ ni]) if fi(t1, ..., tni) ∈ Wτlin(Xn) and
ˆ
α[tj] is already defined for all 1≤j ≤ni.
In [5], the author showed that the set Hyplin(τ) of all linear hypersubstitutions of type τ endowed with a binary operation ◦h, defined by α1 ◦hα2 := ˆα1◦α2, forms a monoid, where αid, defined by αid(fi) := fi(x1, ..., xni), is its neutral element. The algebraic properties Hyplin(τ) := (Hyplin(τ);◦h, αid) is a monoid.
2.4 Coalgebras
Denote the set of all positive integers by N+. That is, N+:={1,2,3, ...}.
Definition 2.4.1. [8] A coalgebra of type τ is a system (A; (fiA)i∈I) consisting of a non-empty set A and a set of finitary co-operations fiA : A → Atni, where ni is the arity of the co-operation fiA and Atni is the ni-th copower of A. We recall that such copowers are defined by Atni := {1, ..., ni} ×A. This means that each ni-ary co-operation fiA is uniquely determined by a pair ((fiA)1,(fiA)2) of mappings, (fiA)1 from A to {1, ..., ni} and (fiA)2 from A to A.
Definition 2.4.2. [8] Let τ = (ni)i∈I be a type of coalgebra, that is, an indexed set of natural numbers, ni ≥ 1, for all i ∈ I, with corresponding indexed set (fi)i∈I of co-operation symbols. We say that symbol fi has arity ni, for i ∈ I. Let S
{enj | n ≥ 1, n ∈ N,0 ≤ j ≤ n−1} be a set of symbols which is disjoint from the set {fi |i∈I}. We assign to each enj the positive integer n as its arity. Then coterms of type τ are defined as follows:
1. For every i∈I the co-operation symbol fi is an ni-ary coterm of type τ.
2. For every n≥1 and 0≤j ≤n−1 the symbol enj is an n-ary coterm of type τ.
3. If t1, ..., tni are n-ary coterms of type τ, then fi[t1, ..., tni] is an n-ary coterm of type τ and if t0, ..., tn−1 are m-ary coterms of type τ, then enj[t0, ..., tn−1] is an m-ary coterm of type τ, for every i∈I and n≥1 and 0≤j ≤n−1.
Let cTτ(n) be the set of all n-ary coterms of type τ and let cTτ := [
n≥1
cTτ(n) be the set of all (finitary) coterms of type τ.
10 Definition 2.4.3. [7] The operation Smn : cTτ(n) ×(cTτ(m))n → cTτ(m) is defined by induction on the complexity of coterm, as follows:
(i) Smn(eni, t0, ..., tn−1) :=ti for 0≤i≤n−1.
(ii) Snni
i(fi, en0i, ..., enni
i−1) := fi for an ni-ary co-operation symbol fi.
(iii) Smnj(gj, t1, ..., tnj) :=gj[t1, ..., tnj] if gj is an nj-ary co-operation symbol.
(iv) Smn(fi[s1, ..., sni], t1, ..., tn) := fi[Smn(s1, t1, ..., tn), ..., Smn(sni, t1, ..., tn)] where fi
is an ni-ary co-operation symbol, s1, ..., sni are n-ary coterms of type τ and t1, ..., tn are m-ary coterms of type τ.
These operations give us a many-sorted algebra
cTτ := ((cTτ(n))n≥1,(Smn)m,n≥1,(enj)0≤j≤n−1).
Theorem 2.4.4. [7] The many-sorted algebra cTτ satisfies the following identities:
(C1) Sˆmp(z,Sˆmn(y1, x1, ..., xn), ...,Sˆmn(yp, x1, ..., xn))≈ Sˆmn( ˆSnp(z, y1, ..., yp), x1, ..., xn), (m, n, p∈N+),
(C2) Sˆmn(eni, x0, ..., xn−1)≈xi (m ∈N+,0≤i≤n−1), (C3) Sˆnn(y, en0, ..., enn−1)≈y, (n∈N+).
(here Sˆmn, eni are operation symbols corresponding to the clone type.)
Let (A; (fiA)i∈I) be a coalgebra of type τ. Then each coterm of type τ induces a cooperation which is inductively defined as follows:
(i) If fi is an ni-ary cooperation symbol, then fiA is the induced ni-ary cooperation.
(ii) (enj)A:=ιn,Aj for every n≥1,0≤j ≤n−1, where ιn,Aj is an n-ary injection.
(iii) If fi[g1, ..., gni] is a coterm and if we inductively assume that the induced cooperations g1A, ..., gAni are known, then (fi[g1, ..., gni])A=fiA[gA1, ..., gnAi].
(iv) If g0A, ..., gn−1A are assumed to be known, then we define (enj[g0, ..., gn−1])A =gAj for 0≤j ≤n−1.
11 Definition 2.4.5. [7] A cohypersubstitution of type τ is a mapping σ:{fi |i∈I} → cTτ from the set of all co-operation symbols to the set of all coterms which preserves the arities. The extension of σ is a mapping σˆ : cTτ → cTτ which is inductively defined by the following steps:
(i) σ[eˆ nj] :=enj for every n ≥1 and 0≤j ≤n−1, (ii) σ[fˆ i] :=σ(fi) for every i∈I,
(iii) σ[fˆ i[t1, ..., tni]] :=Snni(σ(fi),σ[tˆ 1], ...,σ[tˆ ni]) for t1, ..., tni ∈cTτ(n).
Let Cohyp(τ) be the set of all cohypersubstitutions of type τ. On the set Cohyp(τ) of all cohypersubstitutions of type τ we may define a binary operation
◦coh by σ1◦cohσ2 := ˆσ1◦σ2 where ◦ is the usual composition of mappings. Let σid
be the cohypersubstitution defined by σid(fi) := fi for all i ∈ I. Then we have that (Cohyp(τ);◦coh, σid)forms a monoid which is called the monoid of cohypersubstitution of type τ where the cohypersubstitution σid satisfies the equation σˆid[t] = t for all t∈cTτ.
Definition 2.4.6. [12] Let m ∈ N+, a generalized superposition of coterms: Sm : (cTτ)(m+1) →cTτ defined inductively by the following steps:
(i) If t =eni and 0≤i≤m−1, then
Sm(eni, t0, ..., tm−1) = ti, where t0, ..., tm−1 ∈cTτ. (ii) If t =eni and 0< m≤i≤n−1, then
Sm(eni, t0, ..., tm−1) = eni, where t0, ..., tm−1 ∈cTτ. (iii) If t =fi[s1, ..., sni], then
Sm(t, t1, ..., tm) =fi(Sm(s1, t1, ..., tm), ..., Sm(sni, t1, ..., tm)) where Sm(s1, t1, ..., tm), ..., Sm(sni, t1, ..., tm)∈cTτ.
The above definition can be written in the following forms:
(i) If t =eni and 0≤i≤m−1, then
eni[t0, ..., tm−1] =ti, where t0, ..., tm−1 ∈cTτ.
12 (ii) If t =eni and 0< m≤i≤n−1, then
eni[t0, ..., tm−1] =eni, where t0, ..., tm−1 ∈cTτ. (iii) If t =fi[s1, ..., sni], then
(fi[s1, ..., sni])[t1, ..., tm] =fi[s1[t1, ..., tm], ..., sni[t1, ..., tm]]
where s1[t1, ..., tm], ..., sni[t1, ..., tm]∈cTτ.
Definition 2.4.7. [12] A generalized cohypersubstitution of type τ is a mapping σ : {fi|i ∈ I} → cTτ. The extension of σ is a mapping σˆ : cTτ → cTτ which is inductively defined by the following steps:
(i) σ[eˆ nj] :=enj for every n ≥1 and 0≤j ≤n−1, (ii) σ[fˆ i] :=σ(fi) for every i∈I,
(iii) σ[fˆ i[t1, ..., tni]] :=Sni(σ(fi),σ[tˆ 1], ...,σ[tˆ ni]) for t1, ..., tni ∈cTτ(n). Let CohypG(τ) be the set of all generalized cohypersubstitutions of type τ.
In [12], on the set CohypG(τ) of all generalized cohypersubstitutions of type τ we define a function◦CG:CohypG(τ)×CohypG(τ)→CohypG(τ)byσ1◦CGσ2 := ˆσ1◦σ2
for all σ1, σ2 ∈CohypG(τ) where ◦ is the usual composition of mappings. Let σid be the generalized cohypersubstitution defined by σid(fi) := fi for all i∈I.
Proposition 2.4.8. [12] The set CohypG(τ) of all generalized cohypersubstitutions of typeτ is associates with a binary operation ◦CGand the generalized cohypersubstitution σid is an identity of CohypGτ. We obtain a monoid (CohypG(τ);◦CG, σid).
CHAPTER 3
LINEAR COHYPERSUBSTITUTIONS OF TYPE τ = (n)
In this chapter, we will introduce the notions of linear cohypersubstitutions of type τ = (n).
3.1 Monoid of linear cohypersubstitutions of type τ = (n)
In this section, we will introduce the notions of the monoid of linear cohypersubstitutions.
Let t be a coterm of type τ. Then we denotes E(t) :={eni |eni occurring in t and 0≤i≤n−1}.
Definition 3.1.1. An n-ary linear coterm of type τ is defined by induction as follows:
(1) For every i∈I the co-operation symbol f is an n-ary linear coterm of type τ. (2) For every n ≥1 and 0≤j ≤n−1 the symbol enj is an n-ary linear coterm of
type τ.
(3) If t1, ..., tn are n-ary linear coterms of type τ and E(tj)∩E(tk) = ∅ for all 1≤j < k ≤n, then f[t1, ..., tn] is an n-ary linear coterm of type τ.
We denote by cTτlin,(n) the set of all n-ary linear coterms of type τ and cTτlin :=
[
n≥1
cTτlin,(n) the set of all (finitary) linear coterms of type τ.
Example 3.1.2. Letτ = (2).Thenf[e30, e31], f[e31, e30]andf[f[e30, e13], e32]are3-ary linear coterms of type τ.
Proposition 3.1.3. If t ∈cTτlin,(n), s1, ..., sn∈cTτlin,(m), and E(sj)∩E(sk) =∅ for all 1≤j < k≤n, then Smn(t, s1, ..., sn)∈cTτlin,(m).
Proof. We will prove by induction on the complexity of the linear coterm t.
(i) If t =eni for 0≤i≤n−1, then
Smn(eni, s1, ..., sn) =si+1 ∈cTτlin,(m). 13
14 (ii) If t =f, then
Snn(f, s1, ..., sn) = Smn(f[en0, ..., enn−1], s1, ..., sn)
= f[snm(en0, s1, ..., sn), ..., snm(enn−1, s1, ..., sn)]
= f[s1, ..., sn]
∈ cTτlin,(m).
(iii) If t = f[t1, ..., tn] ∈ cTτlin,(n), E(tj)∩E(tk) = ∅ for all 1 ≤ j < k ≤ n and assume that Smn(tj, s1, ..., sn)∈cTτlin,(m), then
Smn(f[t1, ..., tn], s1, ..., sn) = f[Smn(t1, s1, ..., sn), ..., Smn(tn, s1, ..., sn)]
∈ cTτlin,(m).
Therefore, Smn(t, s1, ..., sn)∈cTτlin,(m).
Definition 3.1.4. The partial many-sorted mapping Snm :cTτlin,(n)×(cTτlin,(m))n (→ cTτlin,(m) is defined by
Snm(t, s1, ..., sn) :=
Smn(t, s1, ..., sn) if E(sj)∩E(sk) =∅ for 1≤j < k≤n, not defined otherwise.
For the extension of a cohypersubstitution σ the following holds:
Lemma 3.1.5. For any cohypersubstitution σ and any linear coterms t we have E(t)⊇E(ˆσ[t]).
Proof. Let σ ∈Cohyp(τ) and let t∈cTτlin,(n) be a linear coterm.
We will prove by induction on the complexity of the linear coterm t.
(1) If t =eni where n≥1,0≤i≤n−1, then
E(t) = E(eni) = E(ˆσ[eni]) =E(ˆσ[t]).
(2) If t =f[en0, ..., enn−1], then
E(ˆσ[t]) = E(ˆσ[f[en0, ..., enn−1]])
15
= E(Snn(σ(f),σ[eˆ n0], ...,σ[eˆ nn−1]))
= E(Snn(σ(f), en0, ..., enn−1))
⊆
n−1
[
j=0
E(enj)
= E(f[en0, ..., enn−1])
= E(t).
(3) If t=f[t1, ..., tn] with t1, ..., tn ∈cTτlin,(n), E(tj)∩E(tk) =∅ for 1≤j < k≤n and assume that E(tj)⊇E(ˆσ(tj)) where 1≤j ≤n, then
E(ˆσ[t]) = E(ˆσ[f[t1, ..., tn]])
= E(Snn(σ(f),σ[tˆ 1], ...,σ[tˆ n]))
⊆
n
[
j=1
E(ˆσ[tj])
⊆
n
[
j=1
E(tj)
= E(f[t1, ..., tn])
= E(t).
Therefore E(t)⊇E(ˆσ[t]).
Lemma 3.1.6. Let s, t ∈cTτ and σ ∈ Cohyp(τ). If E(s)∩E(t) = ∅ then E(ˆσ[s])∩ E(ˆσ[t]) =∅.
Proof. Assume that E(s)∩E(t) =∅.
Then ∅ ⊆E(ˆσ[s])∩E(ˆσ[t])⊆E(s)∩E(t) =∅.
Therefore, E(ˆσ[s])∩E(ˆσ[t]) = ∅.
Definition 3.1.7. A linear cohypersubstitution of typeτ is a mapping of cohypersubstitution σ : {f} → cTτlin which preserves the arities. The extension of σ is a mapping ˆ
σ:cTτlin→cTτlin which is inductively defined by the following steps:
(i) σ[eˆ nj] :=enj for every n ≥1 and 0≤j ≤n−1, (ii) σ[fˆ ] :=σ(f),
16 (iii) σ[fˆ [t1, ..., tn]] := Snm(σ(f),σ[tˆ 1], ...,σ[tˆ n]) and assume that σ[tˆ j] is already
defined for all 1≤j ≤n.
Note that: By Lemma 3.1.5. and Lemma 3.1.6., we obtain Snm(σ(f),σ[tˆ 1], ...,σ[tˆ n]) = Smn(σ(f),σ[tˆ 1], ...,σ[tˆ n]).
Let Cohyplin(τ) be the set of all linear cohypersubstitutions of type τ.
Theorem 3.1.8. The structure(Cohyplin(τ);◦coh)is a subsemigroup of(Cohyp(τ);◦coh).
Proof. We will show that σ1◦cohσ2 ∈Cohyplin(τ) for all σ1, σ2 ∈Cohyplin(τ).
Let σ1, σ2 ∈Cohyplin(τ). Then
(σ1◦cohσ2)[f] = (ˆσ1◦σ2)[f]
= ˆσ1[σ2(f)]
∈cTτlin,(n). Hence, σ1◦cohσ2 ∈Cohyplin(τ).
Therefore, Cohyplin(τ) is a subsemigroup of Cohyp(τ).
Let σid be the cohypersubstitution defined by σid(f) := f.
Since f is a linear coterm, then σid is a linear cohypersubstitution.
Lemma 3.1.9. The linear cohypersubstitution σid satisfies the equation σˆid[t] = t for all t ∈cTτlin.
Proof. We will prove by induction on the complexity of the linear coterm t.
(1) If t =eni where n≥1,0≤i≤n−1, then
ˆ
σid[t] = ˆσid[eni] =eni =t.
(2) If t =f, then
ˆ
σid[t] = ˆσid[f] =σid(f) =f =t.
17 (3) If t = f[t1, ..., tn] ∈ cTτlin,(n), E(tj)∩E(tk) = ∅ for all 1 ≤ j < k ≤ n and
assume that σˆid[tj] =tj for all 1≤j ≤n. Then ˆ
σid[t] = ˆσid[f[t1, ..., tn]]
= Snn(σid(f),σˆid[t1], ...,σˆid[tn])
= Snn(f, t1, ..., tn)
= f[t1, ..., tn]
= t.
Therefore σˆid[t] =t for all t∈cTτlin,(n).
Theorem 3.1.10. The structure (Cohyplin(τ);◦coh, σid) is a submonoid of (Cohyp(τ);◦coh, σid).
3.2 Idempotent and some regular elements of linear cohypersubstitutions of type τ = (n)
In 2016, D. Boonchari and K. Saengsura were studied the monoid of cohypersubstitutions of type τ = (n) (see [2]). In this section, we determine idempotent and some regular elements of monoid of linear cohypersubstitutions of type τ = (n). Let S be a semigroup, an element a of S is called idempotent if aa = a, and called regular if there exists x∈S such that axa=a (see [4]). For any σ ∈Cohyplin(τ)and τ = (n), if σ(f) = t, we denote σ by σt. We call the symbol enj the injection symbol, for all 0≤j ≤n−1.
Lemma 3.2.1. Let t ∈ cTτlin,(n), s1, ..., sn ∈ cTτlin,(m) and E(sk) ∩E(sl) = ∅ for k, l ∈ {1, ..., n} and k 6= l. If E(t) = {enj−1|∀j ∈ J,∅ 6= J, J ⊆ {1, ..., n}} and sj =enj−1 for all j ∈J, then Smn(t, s1, ..., sn) = t.
Proof. Assume that E(t) ={enj−1|∀j ∈J,∅ 6=J, J ⊆ {1, ..., n}} and sj =enj−1 for all j ∈J.
We give a proof by induction on the complexity of the linear coterm t.
If t =enj−1 for some j ∈J, then
Snm(t, s1, ..., sn) = Smn(enj−1, s1, ..., sn)
18
= sj
= enj−1
= t.
If t =f, then
Snm(t, s1, ..., sn) = Smn(t, s1, ..., sn)
= Smn(f, s1, ..., sn)
= Smn(f[en0, ..., enn−1], s1, ..., sn)
= f[Smn(en0, s1, ..., sn), ..., Smn(enn−1, s1, ..., sn)]
= f[s1, ..., sn]
= t.
If t =f[t1, ..., tn]∈cTτlin,(n) and assume that Smn(ti, s1, ..., sn) =ti for all i= 1, ..., n.
Then
Snm(t, s1, ..., sn) = Smn(t, s1, ..., sn)
= Smn(f[t1, ..., tn], s1, ..., sn)
= f[Smn(t1, s1, ..., sn), ..., Smn(tn, s1, ..., sn)]
= f[t1, ..., tn]
= t.
Therefore, Smn(t, s1, ..., sn) =t.
Theorem 3.2.2. Let σt ∈Cohyplin(n) and t ∈ cTτlin,(n). Then σt is an idempotent if and only if σˆt[t] =t.
Proof. Assume that σt is an idempotent. We obtain
ˆ
σt[t] = ˆσt[σt(f)] = (σt◦cohσt)(f) = σt(f) = t.
Conversely, assume that σˆt[t] =t. Then
(σt◦cohσt)(f) = ˆσt[σt(f)]
19
= ˆσt[t]
= t
= σt(f).
Therefore, σt is an idempotent.
The next result is a condition for an element of Cohyplin(n) to be idempotent.
Theorem 3.2.3. Let σt ∈ Cohyplin(n). If t = eni for all 0 ≤ i ≤ n − 1 or t = f[en0, ..., enn−1], then σt is an idempotent.
Proof. If t =eni, then
ˆ
σt= ˆσt[eni] =eni =t.
By Theorem 3.2.1., we obtain σt is an idempotent.
If t =f[en0, ..., enn−1], then ˆ
σt[t] = ˆσt[f[en0, ..., enn−1]]
= Snn(σt(f),σˆt[en0], ...,σˆt[enn−1])
= Snn(f[en0, ..., enn−1], en0, ..., enn−1)
= f[Smn(en0, en0, ..., enn−1), ..., Smn(enn−1, en0, ..., enn−1)]
= f[en0, ..., enn−1]
= t,
then, σt is an idempotent.
Remark 3.2.4. σt is not idempotent element if t ∈cTτlin,(n)\ {eni, f|0≤i≤n−1}.
Example 3.2.5. Let σt∈Cohyplin(3) where t=f[e31, e32, e30]is a ternary linear coterm of type τ. Then
ˆ
σt[t] = ˆσt[f[e31, e32, e30]]
= S33(σt(f),σˆt[e31],σˆt[e32],σˆt[e30])
= S33(t, e31, e32, e30)
= S33(f[e31, e32, e30], e31, e32, e30)
20
= f[S33(e31, e31, e32, e30), S33(e32, e31, e32, e30), S33(e30, e31, e32, e30)]
= f[e32, e30, e31] 6= t.
Remark. 1 +n is the number of idempotents of linear cohypersubstitution of type τ = (n).
Now, we characterize all regular elements of Cohyplin(n).
Theorem 3.2.6. Let σt ∈Cohyplin(τ) and t∈ cTτ. Then σt is an regular if and only if σˆt[ˆσs[t]] = t for some σs ∈Cohyplin(τ).
Proof. Assume that σt is an regular.
There exists σs ∈Cohyplin(τ) such that σt(f) = (σt◦cohσs◦cohσt)(f).
Thus
ˆ
σt[ˆσs[t]] = ˆσt[ˆσs[σt(f)]] = (σt◦cohσs◦cohσt)(f) = σt(f) = t.
Therefore, σˆt[ˆσs[t]] = t.
Conversely, assume that σˆt[ˆσs[t]] =t for some σs ∈Cohyplin(n). Then (σt◦cohσs◦cohσt)(f) = ˆσt[ˆσs[σt[f]]
= ˆσt[ˆσs[t]]
=t
=σt(f).
Therefore, σt is an regular.
Theorem 3.2.7. Let σt ∈ Cohyplin(τ). If t = enj for all 0 ≤ j ≤ n−1, then σt is regular.
Proof. Assume that t =enj for all 0≤j ≤n−1. We obtain ˆ
σt[ˆσs[t]] = ˆσt[ˆσs[enj]] = enj =t.
Therefore, σt is regular.