This paper explores the rich combinatorial properties of a small subset of the real line near zero by examining the $wap$-compactification of semitopological semigroups and their ultrafilters near an idempotent.
Contribution
It introduces and proves new locally combinatorial concepts near a virtual idempotent using $wap$-compactification of semitopological semigroups.
Findings
01
Ultrafilters near an idempotent form a compact subsemigroup
02
Rich combinatorial properties are observed near zero on the real line
03
The $wap$-compactification provides a framework for analyzing local combinatorial structures
Abstract
A small part of real line which is very close to zero has rich combinatorial properties. The aim of this paper is to express and then prove some locally combinatorial concepts near a virtual idempotent by considering the wap-compactification of a semitopological semigroup S. The wap−compactification of a semitopological semigroup S, is denoted by Sw, the collection of all ultrafilters near an idempotent η∈Sw forms a compact subsemigroup of βSd, where Sd denotes S as discrete space.
Equations77
a+t∈H∑f(t)∈A
a+t∈H∑f(t)∈A
0+={p∈βSd:∀ϵ>0(0,ϵ)∈p}.
0+={p∈βSd:∀ϵ>0(0,ϵ)∈p}.
J0(S)={p∈0+(S):∀A∈p,AisaJ−setnearzero}.
J0(S)={p∈0+(S):∀A∈p,AisaJ−setnearzero}.
J_{\eta}(\mathbb{N})=\{p\in\beta\mathbb{N}:\forall A\in p,\mbox{ $A$ is $J$-set near $\eta$.}\}.
J_{\eta}(\mathbb{N})=\{p\in\beta\mathbb{N}:\forall A\in p,\mbox{ $A$ is $J$-set near $\eta$.}\}.
M=\{p\in\beta\mathbb{N}:\forall A\in p,\mbox{ $A$ is an additive central set}\}.
M=\{p\in\beta\mathbb{N}:\forall A\in p,\mbox{ $A$ is an additive central set}\}.
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Topology and Set Theory · Mathematical Dynamics and Fractals · semigroups and automata theory
Full text
SOME COMBINATORIAL CONCEPTS NEAR AN IDEMPOTENT
A. Pashapournia* and M. A. Tootkaboni** and D. Ebrahimi Bagha
A small part of real line which is very close to zero has rich combinatorial properties.
The aim of this paper is to express and then prove some locally combinatorial concepts near a virtual idempotent by considering the wap-compactification of a semitopological semigroup S. The wap−compactification of a semitopological semigroup S, is denoted by Sw, the collection of all ultrafilters near an idempotent η∈Sw forms a compact subsemigroup of βSd, where Sd denotes S as discrete space.
Key words and phrases:
wap-compactification, Idempotent, J-set, C-set, Central sets Theorem.
*†*Corresponding Author
2020 Mathematics Subject Classification:
Primary 11B25; 43A55; Secondary 22A15; 54D80
1. Introduction
Furstenberg defined the concept of central subset of the natural numbers N in [9]. Also
he stated that for every finite partition of natural numbers, one of the cells contains a central set. The following theorem is the original Central Sets Theorem:
Theorem 1.1**.**
*([9] Proposition 8.21) Let l∈N and for each i∈{1,2,…,l}, let ⟨yi,n⟩n=1∞ be
a sequence in Z. Let C be a central subset of N. Then there exist sequences ⟨an⟩n=1∞ in
N and ⟨Hn⟩n=1∞ in Pf(N)(=the collection of all nonempty finite subsets of N), such that
(1) for all n, maxHn<minHn+1, and
(2) for all F∈Pf(N) and all i∈{1,2,…,l}, ∑n∈F(an+∑t∈Hnyi,t)∈C.
In [10], Furstenberg and Katznelson developed a technique so it was used later to provide another proof of the above Theorem (See [4]). The following theorem is a new version of Central Sets Theorem assuming that S is a commutative semigroup.
Theorem 1.2**.**
*([12] Theorem 14.8.4)
For a central subset A of a commutative semigroup (S,+), there are functions α:Pf(NS)→S and
H:Pf(NS)→Pf(S) such that
(1) if ∅=F⊊G∈Pf(N, then maxH(F)<minH(G), and
(2) if m∈N, {Xi}i=1m⊆Pf(NS), X1⊊X2⊊⋯⊊Xm,
and fi∈Xi for each i∈{1,2,…,m}, then ∑i=1m(α(Xi)+∑t∈H(Xi)fi(t))∈A.
Each set satisfying the conclusion of the Central Sets Theorem is called C-set. Let A be a subset of
commutative semigroup (S,+), whenever
F∈Pf(NS), there exist a∈S and H∈Pf(N) such that
[TABLE]
for each f∈F, we say that A is a J−set.
Let (S,+) be an arbitrary infinite semigroup. The spectrum of l∞(S) is called the Stone-Cˇech of S and is denoted by βS.
It is the biggest semigroup compactification of S. In other words,
(a) right translation for every x∈βS and left translation for each
s∈S are continuous,
(b) the topological center of βS is a dense subset of βS, and
(c) if (X,ϕ) is a semigroup compactification of S, then there exists an onto homomorphism ψ:βS→X such that
ψ∘ε=ϕ, where ε:S→βS is evaluation map.
For more details of semigroup of compactification see [5].
The subset
L is said to be a left ideal of S if s+l∈L for every s∈S and every l∈L. And L is a minimal left ideal of S if
for each left ideal J of S, whenever J⊆L one gets J=L. The minimal
right ideal is defined analogously. An element x∈S is an idempotent if
x+x=x and E(S) denotes the collection of all idempotents in S. If S is a compact Hausdorff right topological semigroup, then
the minimal ideal, K(S), exists and K(S) is a disjoint union of minimal right ideals as well as a disjoint union of minimal left ideals. Every idempotent in K(S) is called minimal idempotent. Every element of a minimal idempotent is called central set. Also a subset A of S is called central∗ set if it is a member of
every minimal idempotent. See [12].
An ultrafilter p on S=(0,+∞) is called near zero if (0,ε)∈p for all ε>0. Let
[TABLE]
Then (0+,+) is a closed subsemigroup of βSd.
In [11], N. Hindman and I. Leader stated some combinatorial results in real numbers near zero. They introduced also new combinatorial applications of the sets which are central near zero. For more detail see [6, 7].
Let S⊆(0,∞) be an additive dense subsemigroup. The set of all
function f:N→S such that limn→∞f(n)=0 is
denoted by T0. The set A⊆S is a central set near zero if
there exists an idempotent p contained in smallest ideal of 0+(S) with A∈p.
In [3], authors gave definitions of the J-set and the C-set near zero and stated the Central sets Theorem near zero.
As a natural consequence, has been proved
[TABLE]
is a two side ideal of 0+(S), and every element of idempotent of J0(S) is C-set.
Remark 1.3**.**
Since (N,+) is a commutative semigroup, so the set Nw as
wap-compactification of the natural numbers is a commutative compact semitopological semigroup. By Lemma 2.1, there exists
a surjective continuous homomorphism π:βN→Nw such that π(p)⊆p for every
p∈βN and π(K(βN))=K(Nw). By Theorem 21.19 and Corollary 2.40 in [12],
K(Nw) is a compact topological group so ∣E(K(Nw))∣=1. Moreover, if η is the minimal idempotent of Nw, therefore η⊆p for every p∈E(K(βN)), and so E(K(βN))⊆η∗.
The concept of J-set near η is partition regular, see Definition 3.1, and Lemma 4.9. So
[TABLE]
is a nonempty set and by Theorem 4.10 is a two sided ideal. As a consequence, we will have E(K(βN))⊆Jη(N)⊆η∗,
and so every central∗ set is a J-set near η, and so is every central set.
Now let
[TABLE]
Then, according to Theorem 16.24 in [12], we get M=cl(E(K(βN,+))⊆η∗ is a left ideal of (βN,⋅). And so, by Theorem 16.26.1, every central∗ set in (N,⋅)
is a central set in (N,+), i.e. every multiplicative central∗ set is an additive central set, and so is a J-set near η.
Assume that
[TABLE]
Then every combinatorial rich ultrafilters belong to η∗, where combinatorial rich ultrafilters is a multiplicative idempotent p∈M∩Δ∩K(βN,⋅). By Theorem 17.1 in [12], there is a combinatorial rich ultrafilter. For some properties of combinatorial rich ultrafilters see Theorem 17.3 in [12]. So it seems that η∗ has rich combinatorial properties.
In this paper, we just concentrate on extension of concepts J-set and C-set near an idempotent in wap-compactification of a semitopological semigroup.
As a result of the above points, we have two versions of the Central sets Theorem on (0,+∞),
Global Central sets Theorem and Local Central sets Theorem. The zero for the semigroup of (0,+∞) is a virtual
idempotent, but it can easily be used for other semigroup. It seems that, the idempotents in the
wap-compactification of a semitopological semigroup can act as a virtual idempotent.
Therefore, the question arises as to whether like the virtual idempotent of [math], the Local Central sets Theorem and related
concepts can be defined for the virtual idempotent of a subsemigroup.
In what follows, we give and prove some locally combinatorial concepts near a virtual idempotent. For
this purpose, we consider the wap-compactification of a semitopological semigroup S.
When S is a discrete semigroup, the Lmc-compactification and βS are topologically isomorphism, and also if S
is a commutative semigroup then wap-compactification of S is commutative, so we fucus on
wap-compactification of a semitopological semigroup S.
According to [1], the Lmc-compactification is expressed as a space of e-ultrafilters.
However, in similar ways, we can describe the wap-compact-ification of a semitopological semigroup as
a space of e-ultrafilters. For this purpose, in brief, for a semitoplogical semigroup of S in Section 2, the wap-compactification of
S is described as a space of e-ultrafilters, which in continues play an essential role. In Section 3, we state some combinatorial concepts
near an idempotent in wap−compactification of a commutative semigroup S and the next section, we concentrate on noncommutative semigroup.
2. Preliminary
Let (S,+) be a semitopological semigroup (i.e., right and left translations are continuous).
The collection of all bounded complex valued continuous functions on S with uniform norm is a C∗-algebra
and is denoted by CB(S). A weakly almost periodic function, f, is a member of
CB(S) such that {Rsf:s∈S} is weakly
relatively compact subset of CB(S). wap(S) denotes the weakly almost periodic functions on S. (ε,Sw) as the wap-compactification of a semitopological semigroup is the universal semitopological semigroup compactification of S, see [5].
In [1], Lmc-compactification has been characterized as a space
of e-ultarfilters. We know that the Lmc-compactification of a discrete semigroup is same the Stone-Cˇech compactification. We restate some definitions and concepts from [1], because we need to describe wap-compactification as a space of e-ultrafilters.
For every complex valued function f on S, Z(f)={s∈S:f(s)=0} is called a zero set. Let Z(wap(S))={Z(f):f∈wap(S)}.
We say that A⊆Z(wap(S)) is a z-filter on wap(S) (z−filter) if,
(i)∅∈/A,
(ii) for every A,B∈A, implies that A⋂B∈A,
(iii) if A∈A, B∈Z(wap(S)) and A⊆B then B∈A.
For a complex valued function f and ϵ>0, define
Eϵ(f)={s∈S:∣f(s)∣≤ϵ}.
For I as a collection of complex valued functions on S, define E(I)={Eϵ(f):f∈I,ϵ>0}. Also, for any
collection A⊆Z(wap(S), let
[TABLE]
If A is a z−filter, we say that A is an e−filter if
E(E−(A))=A. For every z−ultrafilter A, E(E−(A)) is an e−ultrafilter.
For a Hausdorff semitopological semigroup S. By Theorem 3.8 in [1], it is proved that
{p:p\mboxisane−\mboxultrafilter.} and Sw are topologically isomorphic. Therefore,
[TABLE]
forms a basis for a topology on Sw. Each a∈S, define e(a)={Eϵ(f):f(a)=0,ϵ>0} is an e-ultrafilter. Also, for
A∈Z(wap(S)) and x∈S, we have A∈x+p if and only
if λx−1(A)∈p, see Lemma 3.9 in [1].
Now, we ready define ultrafilters near an idempotent. For η∈Sw, we define
[TABLE]
We know that if p,q∈βS and A⊆S, then we have A∈p+q if and only if
{x∈S:−x+A∈q}∈p, where −x+A={y∈S:x+y∈A}. See [12].
Lemma 2.1**.**
*Let (S,+) be a semitopological semigroup.
(a) For η∈Sw, η∗={p∈βSd:η⊆p}.
(b) Let A⊆S. Then x∈clSwA if and only if clβSdA∩x∗=∅.
(c) For each u,v∈Sw, u∗+v∗⊆(u+v)∗.
(d) Let η∈E(S∗), then η∗ is a compact subsemigroup of βSd.*
Proof.
(a) It is obvious.
(b) Let x∈clSwA, so for each U∈x we will have U∩A=∅. Therefore x∪{A}
has the finite intersection property, this implies that there exists an ultrafilter p such that x∪{A}⊆p.
So clβSdA∩x∗=∅.
Conversely, let p∈clβSdA∩x∗. therefore A∈p and x⊆p.
This implies that for each U∈x we will have U∩A=∅. So x∈clSwA.
(c) Let p∈x∗ and q∈y∗. We prove p+q∈(x+y)∗. In order to we show that
x+y⊆p+q. So let U∈x+y, then there are ϵ>0 and f in wap(S) such that
U=Eϵ(f) and Eδ(y,f)={t∈S:λt−1(Eδ(f))∈y}∈x for each δ>0. Now let δ=ϵ, so
[TABLE]
This implies that U=Eϵ(f)∈p+q.
(d) It is obvious that η∗ is a subsemigroup of βSd. Now let p∈βSd∖η∗.
Pick U∈η∖p. Then clβSd(S∖U) is a neighborhood of p which
[TABLE]
Therefore η∗ is closed subset of βSd.
∎
Lemma 2.2**.**
*Let (S,+) be a semitopological semigroup.
a) For each p∈βSd, there exists a unique ep∈Sw such that ep⊆p.
b) π:βSd→Sw by π(p)=ep is well defined.
c) π:βSd→Sw is continuous homomorphism. In particular, for p,q∈βSd, π(p)+π(q)⊆p+q.*
Proof.
It is obvious.
∎
Definition 2.3**.**
Let (S,+) be a semitopological semigroup and η∈E(S∗). A subset A of S is piecewise syndetic near
η if and only if K(η∗)∩clβS(A)=∅.
3. ** J-sets and C-sets near an idempotent**
Let (S,+) be a commutative semitopological semigroup. We define concepts of J-set and
C-set near a virtual idempotent η in Sw.
For B⊆N, the upper density of B is defined by
[TABLE]
Let η∈Sw and f be a sequence in S.
We say that d−limn∈Nf(n)=η if
for every U∈η, d({n:f(n)∈/U})=0. In this paper,
[TABLE]
Let f,g∈Tx, and set A={n∈N:f(n)∈/U} and B={n∈N:g(n)∈/U}.
Then d(A)=d(B)=0, therefore d(A∪B)=0. By Exercise 3.11(c) in [13],
d(A∪B)=1−d((A∪B)c) implies that d(Ac∩Bc)=1. So
d(f−1(U)∩g−1(U))=1 for each U∈x.
Definition 3.1**.**
Let A⊆S and x∈clSwA. A⊆S is called J−set near x if for every
F∈Pf(Tx) and for each U∈x there are a∈U and
H∈Pf(N) such that
[TABLE]
for each f∈F.
Lemma 3.2**.**
If η∈E(S∗) and W∈η, then W is a J-set near η.
Proof.
Pick k∈N and let F={f1,⋯,fl}∈Pf(Tη).
Since
[TABLE]
for each V∈η, so
by definition of Tη, A=⋂g∈Fg(⋂f∈Ff−1(V))
is infinite. Now pick p1,⋯,pn∈clβSdA∩η∗. By Theorem 1.6.13 in [8], so there exist nets
[TABLE]
in N and {aU}U∈η such that aU∈UaU→η and
fi(γαj)→η for each i∈{1,⋯,l} and j∈{1,⋯,l} in Sw. Then
[TABLE]
in ×i=1lSw.
Now for W∈η, ×i=1lW∈×i=1lη. So there exist
H∈Pf(N) and aU∈U⊆W such that
[TABLE]
This implies that W is a J-set.
∎
Lemma 3.3**.**
Let A be a subset of commutative semigroup S, and η∈E(S∗). Then for every m∈N, every
F∈Pf(Tη), and for each U∈η there are a∈U and
H∈Pf(N) such that minH>m and a+∑t∈Hf(t)∈A for each f∈F.
For commutative semitopological semigroup S, let A⊆S.
Then every piecewise syndetic set near η is a J-set near η.
Proof.
Let A⊆S be a piecewise syndetic set near η, and let F={f1,⋯,fl}∈Pf(Tη).
Let Y=×t=1lη∗⊆×t=1lβSd.
So Y is a compact right topological semigroup and for every s∈×t=1lS,
λs is continuous, see Theorem 2.22 in [12]. Pick i∈N and U∈η. Define
[TABLE]
and let Ei,U=Ii,U∪{(a,⋯,a):a∈U}.
Let E=⋂i∈N,U∈eEi,U and let I=⋂i∈N,U∈eIi,U.
Since Ii,U⊆Ei,U⊆×t=1lU for each U∈η,
E⊆Y and I⊆E.
Now pick p,q∈E. We prove
p+q∈E, (i.e. E is subsemigroup). Also, p+q∈I if p∈I or q∈I, (i.e. I is two-side ideal). For U∈η,
so p+q is interior point of W=clβSdU. Pick i∈N, since right translation ρq is continuous, there exists
open set V such that p∈V and
V+q⊆W. Now choose x∈Ei,U∩V with x∈Ii,U if p∈I. Then for some a∈U and H∈Pf(N) with minH>i, x=(a+∑t∈Hf1(t),⋯,a+∑t∈Hfl(t)), whenever x∈Ii,U. Now pick j=maxH. Otherwise, let j=i. So choose an open set Q such that q∈Q and x+Q⊆W, because left translatio λx is continuous. Pick y∈Ej,U∩W with y∈Ij,U if q∈I.
Then x+y∈Ei,U∩W and if either p∈I or q∈I, then x+y∈Ii,U∩W.
Since K(Y)=×t=1lK(η∗), see Theorem 2.23 in [12]. Choose p∈K(η∗)∩A.
So p=(p,⋯,p)∈K(Y). We show that p∈E. Therefore, let Z be an open neighborhood of
p, let i∈N, and choose C1,⋯,Cl∈p such that ×t=1lCt⊆Z. For
a∈⋂t=1lCt. Then a=(a,⋯,a)∈W∩Ei,U. Thus
p∈K(Y)∩E and consequently K(Y)∩E=∅. So we have that
K(E)=K(Y)∩E and so p∈K(E)⊆I. Therefore I1,U∩×t=1lA=∅ for each
U∈η, so choose z∈I1,U∩×t=1lA and a∈U and H∈Pf(N) such that
[TABLE]
∎
Theorem 3.5**.**
Let S be a commutative semitopological semigroup and η be a idempotent in S∗. Let A be a central subset of S near η. Then for U∈η,
there exist functions αU:Pf(Tη)→S and HU:Pf(Tη)→Pf(N) such that
(1)* αU(F)∈U for each F∈Pf(Tη),*
(2)* for F,G∈Pf(Tη) and F⊊G, implies that maxHU(F)<minHU(G), and*
(3)* whenever m∈N, {Xi}i=1m⊆Pf(Tη), X1⊊X2⊂⋯⊊Xm, and
for every i∈{1,2,⋯,m}, fi∈Xi, we have*
[TABLE]
Proof.
Choose a minimal idempotent p of η∗ such that A∈p. Let A∗={x∈A:−x+A∈p}, so A∗∈p.
Therefore, x∈A∗, implies that −x+A∗∈p, see Lemma 4.14 in [12].
We define αU(F)∈S and HU(F)∈Pf(N) for F∈Pf(Tη) and U∈η. By induction on ∣F∣ satisfying the following statements:
(1) αU(G)∈U for each G∈η,
(2) if F,G∈Pf(Tη) and F⊊G, then maxHU(F)<minHU(G), and
(3) whenever m∈N, {Xi}i=1m⊆Pf(Tη), X1⊊X2⊂⋯⊊Xm, and
for each i∈{1,2,⋯,m}, fi∈Xi, one has
[TABLE]
Assume that F={f}. As A∗ is piecewise syndetic near η, choose for U∈η, a∈S∩U and L∈Pf(N) such that
a+∑t∈Lf(t)∈A∗. Define αU({f})=a and
HU({f})=L.
Assume that ∣F∣>1, for every proper subsets G of F and for each U∈η, αU(G) and HU(G) have been defined.
For U∈η, define KU=⋃{HU(G):∅=G⊊F} and let m=maxKU. Define
[TABLE]
Therefore MU is finite set and by (3), MU⊆A∗. Assume that
B=A∗∩⋂x∈MU(−x+A∗), so B∈p and so pick a∈S∩U and L∈Pf(N) such that a+∑t∈Lf(t)∈B for each f∈F.
Let αU(F)=a and HU(F)=L.
The hypothesis (1) is clear. Since minL>m, the hypothesis (2) is satisfied. To verify (3), choose U∈η and
n∈N, let ∅⊊X1⊂⋯⊊Xn=F, and {fi}i=1n∈×i=1nXi.
If n=1, then αU(X1)+∑t∈HU(X1)f1(t)=a+∑t∈Lf1(t)∈B⊆A∗. So let n>1 and
define y=\sum_{i=1}^{n-1}\big{(}\alpha_{U}(X_{i})+\sum_{t\in H_{U}(X_{i})}f_{i}(t)\big{)}. Therefore
y∈MU so a+∑t∈Lf1(t)∈B⊆(−y+A∗) and thus
\sum_{i=1}^{n}\big{(}\alpha_{U}(X_{i})+\sum_{t\in H_{U}(X_{i})}f_{i}(t)\big{)}=y+a+\sum_{t\in L}f_{1}(t)\in A^{*} as required.
∎
Definition 3.6**.**
For a commutative semitopological semigroup S let η∈E(S∗). A⊆S is called C-set near η if
for U∈η,
there are functions αU:Pf(Tη)→S and HU:Pf(Tη)→Pf(N) such that
(1) αU(F)∈U, for each F∈Pf(Tη),
(2) if F,G∈Pf(Tη) and F⊊G, then maxHU(F)<minHU(G) and
(3) whenever m∈N, {Xi}i=1m⊆Pf(Tη), X1⊊X2⊂⋯⊊Xm, and
for each i∈{1,2,⋯,m}, fi∈Xi, one has
[TABLE]
4. **Noncommutative version **
In this section, we assume that (S,+) is noncommutative semigroup. We state the concepts of J-set and C-set near an idempotent. Define
[TABLE]
Definition 4.1**.**
For an arbitrary semitopological semigroup S and for m∈N, we define
[TABLE]
[TABLE]
and
[TABLE]
for U∈η.
Definition 4.2**.**
For an arbitrary semitopological semigroup S and for η∈E(S∗).
Given m∈N, U∈η, a∈SUm+1, t∈Jm, and f∈Tη, define
[TABLE]
Definition 4.3**.**
For an arbitrary semitopological semigroup S, η∈E(S∗), A⊆S, and η∈clSwA.
(a) A is called J-set near η if for each F∈Pf(Tη) and for each U∈η there
exist m∈N, a∈SUm+1, and t∈Jm such that for each f∈F, x(m,a,t,f)∈A.
b) Jη(S)={p∈η∗:∀A∈p,A\mboxisaJ−\mboxsetnearη}.
Lemma 4.4**.**
For a commutative semitopological semigroup S, η∈E(S∗), A⊆S, and η∈clSwA. Then the follwing statemets equivalent:
If (a) is true. Pick F∈Pf(Tη), U∈η, c∈N and for
f∈F, define gf∈Tη by gf(n)=f(n+c). Pick b∈S and
H∈Pf(N) such that for each f∈F, b+∑t∈Hgf(t)∈A. Let m=∣H∣, and let t=\big{(}t(1),\cdots,t(m)\big{)} enumerate H in increasing order. Define a(1)=b
and for j∈{2,⋯,m+1}, define a(j)=c. This complete the proof.
∎
Definition 4.5**.**
For an arbitrary semitopological semigroup S, η∈E(S∗), A⊆S, and η∈clSwA. We say
A is C-set near η if for each U∈η, there are mU:Pf(Tη)→N,
αU∈×F∈Pf(Tη)SUmU(F)+1,
and τU∈×F∈Pf(Tη)JmU(F) such that
(1) if F,G∈Pf(Tη) and F⊊G then τU(F)(mU(F))<τU(G)(1) for each U∈η, and
(2) whenever n∈N, X1,⋯,Xn∈Pf(Tη),
X1⊊X2⊂⋯⊊Xn, and for each
i∈{1,⋯,n}, fi∈Xi, one has
[TABLE]
Lemma 4.6**.**
Let S be a commutative semitopological semigroup and η∈E(S∗),
let A⊆S, and η∈clSwA. Then Definitions 3.6 and 4.5 are equivalent.
Proof.
It is obvious that Definition 4.5 implies Definition 3.6.
Now let Definition 3.6 be true. Pick αU:Pf(Tη)→S and
HU:Pf(Tη)→Pf(N) for each U∈η
as guaranteed by (1) and (2). Now pick c∈N and for f∈Tη define gf∈Tη by
gf(s)=f(s+c), for s∈N. For F∈Pf(Tη), we define
inductively on ∣F∣ a set K(F)∈Pf(Tη) such that
(1){gf:f∈F}⊆K(F) and
(2) if ∅=G⊊F, then K(G)⊊K(F).
If F={f}, let K(F)={gf}. Now let ∣F∣>1 and K(G) has been defined for all proper nonempty
subsets of F. Pick
[TABLE]
and
let
[TABLE]
Now for each U∈η, we define
mU:Pf(Tη)→N, αU′∈×F∈Pf(Tη)SmU(F)+1,
and τU∈×F∈Pf(Tη)JmU(F).
Let F∈Pf(Tη) be given and let mU(F)=∣HU(K(F))∣. Define
αU′(F)∈SmU(F)+1, for j∈{1,2,⋯,mU(F)+1}, αU′(F)(j)=αU(K(F)) if j=1 and
αU′(F)(j)=c if j>1. Let τU(F)=(τU(F)(1),⋯,τ(F)(mU(F))) enumerate HU(K(F)) in increasing order. We need to show that
(a) if F,G∈Pf(Tη) and F⊊G, then τU(F)(mU(F))<τU(G)(1) for each U∈η, and
(b) whenever n∈N, X1,⋯,Xn∈Pf(Tη),
X1⊊X2⊂⋯⊊Xn, and for each
i∈{1,⋯,n}, fi∈Xi, one has
[TABLE]
To verify (a), let F,G∈Pf(Tη) with
F⊊G, then K(F)⊊K(F), and so
[TABLE]
To verify (b), let n∈N, X1,⋯,Xn∈Pf(Tη),
X1⊊X2⊂⋯⊊Xn, and for each
i∈{1,⋯,n}, let fi∈Xi. Then K(X1)⊂K(X2)⊂⋯⊊K(Xn), and for each fi∈Xi,
gfi∈K(Xi) so
∑i=1n(αU(K(Xi)))+∑t∈HU(K(Xi))gfi(t))∈A and
[TABLE]
∎
Lemma 4.7**.**
Let S be a commutative semitopological semigroup, η∈E(S∗), A⊆S, and η∈clSwA. Let A be a J-set near η in S, then for each
F∈Pf(Tη), each U∈η and each n∈N, there exist m∈N, a∈SUm+1, and y∈Jm such that y(1)>n and for each f∈F, x(m,a,y,f)∈A.
Proof.
Pick F∈Pf(Tη), U∈e and n∈N. For each f∈F define Xf∈Tη for
u∈N, by Xf(u)=f(u+n). Pick m∈N, a∈SUm+1 and t∈Jm such that
for each f∈F, x(m,a,t,Xf)∈A. Define y∈Jm by y(i)=n+t(i) for i∈{1,2,⋯,m}.
Then y(1)>1 and for each f∈F, x(m,a,y,f)∈A.
∎
Lemma 4.8**.**
For an arbitrary semitopological semigroup S, η∈E(S∗), A⊆S, and η∈clSwA. Pick U∈η, and let
m,r∈N, let a∈SUm+1,
let t∈Jm, and for each y∈N, let cy∈SUr+1 and zy∈Jr be a such that for each y∈N, zy(r)<zy+1(1). Then there exist
u∈N, d∈SUu+1, and q∈Ju such that for
each f∈Tη,
Let S be a semitopological semigroup, η∈E(S∗), A⊆S, and η∈clSwA, A⊆S, and η∈clSwA. Then A∩Jη(S)=∅ if and only if A is a J-set near η.
Proof.
The necessity is trivial. By Lemma 4.9, J-sets are partition regular.
So, if A is a J-set near η, by Theorem 3.11 in [12] , there is some p∈βS
such that A∈p and for every B∈p, B is a J-set near η.
∎
Corollary 4.12**.**
Let S be a semitopological semigroup, η∈E(S∗), A⊆S, η∈clSwA, and A be a piecewise syndetic near η
subset of S. Then A is a J-set near η.
Proof.
Since A is piecewise syndetic near η, A∩K(η∗)=∅. Therefore K(η∗)⊆Jη(S) implies that
A∩Jη(S)=∅ so by Theorem 4.11, A is a J-set near η.
∎
Theorem 4.13**.**
Let S be a semitopological semigroup, η∈E(S∗), A⊆S, and η∈clSwA. If there is an idempotent in A∩Jη(S), then A is a C-set near η.
Let S be a semitopological semigroup, η∈E(S∗), A⊆S, η∈clSwA, and A⊆S be a central set near η in S. Then A is
a C-set near η.
Proof.
It is obvious.
∎
Lemma 4.15**.**
Let R be a set, let (D,≤) be a directed set, and let S be a semitopological subsemigroup of (T,+).
Let {Ti}i∈D be a decreasing family of nonempty subsets of S such that
1)* η∈clTwTi,*
2)* ⋂i∈DTi=∅, and*
3)* for each i∈D and each x∈Ti there is some j∈D such that x+Tj⊆Ti.*
Let
T=⋂i∈DclβSdTi. Then T is a compact subsemigroup of η∗(S). Let {Ei}i∈D and
{Ii}i∈D be decreasing families of nonempty subsets of ×t∈JS with the following
properties:
(a)* for each i∈D, Ii⊆Ei⊆×t∈JTi,*
(b)* for each i∈D and each x∈Ii there exists j∈D such that x+Ej⊆Ii,
and*
(c)* for each i∈D and each x∈Ei∖Ii there exists j∈D such that
x+Ej⊆Ei and x+Ij⊆Ii.*
Let Y=×t∈Jη∗(S), let E=⋂i∈DclYEi, and let I=⋂i∈DclYIi. Then E is
a subsemigroup of ×t∈JT and I is an ideal of E. If, in addition, either
(d)* for each i∈D, Ti=S and {a∈S:a∈/Ei} is not piecewise syndetic near η, or*
(e)* for each i∈D and each a∈Ti , a∈Ei,*
then given
any p∈K(T), one has p∈E∩K(×t∈JT)=K(E)⊆I.
Let S be a semitopological semigroup, η∈E(S∗), A⊆S, and η∈clSwA. Then A is a C-set near η
if and only if there is an idempotent in A∩Jη(S).
Proof.
The proof is similar to Theorem 14.15.1 in [12].
∎
Bibliography13
The reference list from the paper itself. Each links out to its DOI / PubMed record.
1[1] A. Akbari Tootkaboni, Lmc-compactification of a semitopological semigroup as a space of e-ultrafilters, New York J. Math 19 (2013), 669-688.
2[2] A. Akbari Tootkaboni and T. Vahed The semigroup of ultrafilters near an idempotent of a semitopological semigroup, Topology and its Application, 159 (2012), 3494-3503.
3[3] E. Bayatmanesh and A. Akbari Tootkaboni, Central sets theorem near zero, Topology and its Application, 210 (2016), 70-80.
4[4] V. Bergelson and N. Hindman, Nonmetrizable topological dynamics and Ramsey Theory , Trans. Amer. Math. Soc. 320 (1990), 293-320.
5[5] J. F. Berglund, H. D. Junghenn and P. Milnes, Analysis on Semigroups: Function Spaces, Compactifications, Representations, Wiley, New York., 1989.
6[6] D. De and N. Hindman, Image Partition Regularity Near Zero, Discrete Math. 309 (2009), 3219-3232.
7[7] D. De, N. Hindman, and D. Strauss, Sets central with respect to certain subsemigroup of β S d 𝛽 subscript 𝑆 𝑑 \beta S_{d} , Topology Proceedings, 33 (2009),55-79.
8[8] R. Engelking, General Topology, Berlin: Heldermann, 1989.