• No results found

The binomial sequence spaces of nonabsolute type

N/A
N/A
Protected

Academic year: 2020

Share "The binomial sequence spaces of nonabsolute type"

Copied!
16
0
0

Loading.... (view fulltext now)

Full text

(1)

R E S E A R C H

Open Access

The binomial sequence spaces of

nonabsolute type

Mustafa Cemil Bi¸sgin

*

*Correspondence:

[email protected] Faculty Of Arts And Sciences, Department Of Mathematics, Recep Tayyip Erdo ˘gan University, Zihni Derin Campus, Rize, 53100, Turkey

Abstract

In this paper, we introduce the binomial sequence spacesbr0,sandbr,s

c of nonabsolute

type which include the spacesc0andc, respectively. Also, we prove that the spaces

br0,sandbr,s

c are linearly isomorphic to the spacesc0andc, in turn, and we investigate some inclusion relations. Moreover, we obtain the Schauder bases of those spaces and determine their

α

-,

β

-, and

γ

-duals. Finally, we characterize some matrix classes related to those spaces.

MSC: 40C05; 40H05; 46B45

Keywords: matrix transformations; matrix domain; Schauder basis;

α

-,

β

- and

γ

-duals; matrix classes

1 The basic definitions and notations

Letwbe the set of all real (or complex) valued sequences. Thenwbecomes a vector space under point-wise addition and scalar multiplication. A sequence space is a vector sub-space ofw. We use the notations of,c,c, andp for the spaces of all bounded, null,

convergent, and absolutelyp-summable sequences, respectively, where p<∞. A Banach sequence space is called aBK-space provided each of the mapspn:X−→C

defined bypn(x) =xnis continuous for alln∈N[]. By taking into account the definition

above, one can say that the sequence spaces,c, andcareBK-spaces with their usual sup-norm defined byx∞=supk∈N|xk|andpis aBK-space with itsp-norm defined by

xp=

k= |xk|p

p ,

wherep∈[,∞).

LetA= (ank) be an infinite matrix with complex entries andxw, then theA-transform

ofxis defined by

(Ax)n=

k=

ankxk (.)

and is assumed to be convergent for alln∈N[]. For brevity in the notation, we henceforth prefer that the summation without limits runs from  to∞.

(2)

LetXandYbe two arbitrary sequence spaces andA= (ank) be an infinite matrix. Then

the domain ofAis denoted byXAand defined by

XA=

x= (xk)∈w:AxX

, (.)

which is also a sequence space and the class of all matricesAsuch thatXYAis denoted

by (X:Y). Moreover,A= (ank) is called conservative ifccA. Furthermore,A= (ank)

is calledm-multiplicative, iflimn→∞(Ax)n=mlimn→∞xnfor allxcand the class of all

m-multiplicative matrices is denoted by (X:Y)m. Specially,A= (ank) is called regular, if

A= (ank) is -multiplicative.

The spaces of all bounded and convergent series are denoted bybsandcsand are defined by aid of the matrix domain of the summation matrixS= (snk) such thatbs= (∞)Sand

cs=cS, respectively, whereS= (snk) is defined by

snk=

⎧ ⎨ ⎩

, ≤kn,

, k>n,

for allk,n∈N. A matrixA= (ank) is said to be a triangle ifank=  fork>nandann=  for

alln,k∈N. Furthermore, a triangle matrix uniquely has an inverse, which is also a triangle matrix.

The theory of matrix transformations is of great importance in the summability which was obtained by Cesàro, Borel, Riesz and others. Therefore, many authors have defined new sequence spaces by using this theory. For example, ()Nq andcNq in [],XpandX∞ in [],c˜andc˜ in [],ar andarc in []. Moreover, many authors have constructed new

sequence spaces by using especially the Euler matrix. For instance,erander

cin [],erpand

er

∞ in [] and [],er(),erc(), ander∞() in [],er((m)),erc((m)) ander∞((m)) in [],

er

(B(m)),erc(B(m)), ander∞(B(m)) in [],er(,p),erc(,p), ander∞(,p) in [],er(u,p) and er

c(u,p) in [].

In this paper, we introduce the binomial sequence spaces br,sandbr,s

c of nonabsolute

type which include the spacescandc, respectively. Also, we prove that the spacesbr,sand br,s

c are linearly isomorphic to the spacescandc, in turn and investigate some inclusion

relations. Moreover, we obtain the Schauder basis of those spaces and determine theirα-,

β-, andγ-duals. Finally, we characterize some matrix classes related to those spaces.

2 The binomial sequence spaces of nonabsolute type

In this chapter, we introduce the binomial sequence spacesbr,s andbr,s

c of nonabsolute

type and prove that the spacesbr,sandbr,s

c are linearly isomorphic to the spacescandc, respectively. Moreover, we deal with an inclusion relation related to those spaces.

Letr,s∈Randr+s= . Then the binomial matrixBr,s= (br,s

nk) is defined by

brnk,s= ⎧ ⎨ ⎩

 (s+r)n

n

k

snkrk, ≤kn,

, k>n,

(3)

(i) supnNnk=|(s+r)n n

k

snkrk|<∞,

(ii) limn→∞(s+r)n n

k

snkrk= ,

(iii) limn→∞nk=  (s+r)n

n

k

snkrk= .

Therefore, the binomial matrixBr,s= (brnk,s) is regular forsr> . Unless stated otherwise, we assume thatsr> .

Here, we would like to emphasize that if we taker+s= , we obtain the Euler matrix Er= (er

nk). So the binomial matrixBr,s= (b r,s

nk) generalizes the Euler matrixEr= (ernk).

By considering the definition of the binomial matrixBr,s= (br,s

nk), we define the binomial

sequence spacesbr,sandbr,s

c as follows:

br,s=

x= (xk)∈w: lim n→∞

 (s+r)n

n

k=

n k

snkrkxk= 

and

brc,s=

x= (xk)∈w: lim n→∞

 (s+r)n

n

k=

n k

snkrkxkexists

.

The sequence spacesbr,sandbr,s

c can be redefined by using the notion of (.) as follows:

br,s= (c)Br,s and br,s

c =cBr,s. (.)

It is clear thatbr,sbr,s

c . Letx= (xk) be an arbitrary sequence. Then theBr,s-transform of

x= (xk) is defined by

Br,sxk=yk=

 (s+r)k

k

j=

k j

skjrjxj (.)

for allk∈N.

Now, we want to start with the following theorem related to the theory ofBK-spaces, which is of great importance in the characterization of matrix transformations between sequence spaces.

Theorem . The binomial sequence spaces br,s and br,s

c are BK -spaces with their

sup-norms defined by

xbr,s=xbrc,s=B

r,sx

∞=sup

n∈N

Br,sxn.

Proof The sequence spacescandcareBK-spaces according to theirsup-norms. More-over, the binomial matrixBr,s= (br,s

nk) is a triangle matrix and (.) holds. By combining

these three facts and Theorem .. of Wilansky [], we deduce that the binomial se-quence spacesbr,sandbr,s

c areBK-spaces. This completes the proof.

Let|x|= (|xk|) for allk∈N. Because ofxbr,s=|x|br,sandxbcr,s=|x|brc,sfor at least one sequence in the binomial sequence spacesbr,sandbrc,s,b

r,s

 andbrc,sare sequence spaces

(4)

Theorem . The binomial sequence spaces br,sand br,s

c are linearly isomorphic to the

sequence spaces cand c,respectively.

Proof Because a repetition of similar statements is redundant, the proof of the theorem is given for only the spacebr,s

c . For this purpose, we should show the existence of a linear

bijection between the spacesbrc,sandc. Let us consider the transformationLdefined by L:brc,s−→c,L(x) =Br,sx. Then it is obvious that, for allxbrc,s,L(x) =Br,sxc. Moreover, it is clear thatLis a linear transformation andx=  wheneverL(x) = . Because of this,L is injective.

Lety= (yk)∈cbe given. We define a sequencex= (xk) by

xk=

rk

k

j=

k j

(–s)kj(s+r)jyj

for allk∈N. Then we obtain

Br,sxn=  (s+r)n

n

k=

n k

snkrkxk

= 

(s+r)n n

k=

n k

snk

k

j=

k j

(–s)kj(s+r)jyj

=yn

for alln∈N. So,Br,sx=yand sinceyc, we conclude thatBr,sxc. Hence, we deduce

thatxbr,s

c andL(x) =y. On account of thisLis surjective.

Moreover, we have for allxbr,s c

L(x)

∞=Br,sx∞=xbrc,s.

SoLis norm preserving. Consequently,Lis a linear bijection. Then we obtain the fact that the spacesbr,s

c andcare linearly isomorphic, that is,brc,s∼=c. This completes the proof.

Theorem . The inclusions er

⊂b

r,s

and ercbcr,sstrictly hold,where erand ercare Euler

sequence spaces of nonabsolute type.

Proof Ifr+s= , we obtainEr=Br,s. So, the inclusioner

⊂br,sholds. Assume that  <r<  ands= . Now, we define a sequencex= (xk) such thatxk= (–r)kfor allk∈N. Then it is

obvious thatx= ((–r)k) /∈candErx= ((– –r)k) /∈c. On the other hand,Br,sx= ((+r)k)∈ c. As a consequence,x= (xk)∈br,s\er.

This shows that the inclusioner

⊂br,sstrictly holds. Another part of the theorem can be proved in a similar way. This completes the proof.

Theorem . The inclusions c⊂br,sand cbrc,sstrictly hold.But the sequence spaces

br,sanddo not include each other.

Proof If we consider regularity of the binomial matrixBr,s, we can easily conclude that Br,sxcwheneverxc. This means thatxbr,s

 for allxc, namelycb

r,s

(5)

we define a sequenceu= (uk) such thatuk= (–)kfor allk∈N. Then we obtainBr,sx=

((sr s+r)

k)c

. As a consequence,uis inbr,sbut not inc. So, the inclusionc⊂br,sis strict. By using a similar way, one can show that the inclusioncbr,s

c is strict.

To prove the second part of the theorem, we consider the sequencese= (, , , . . .) and v= (vk) defined byvk= (–sr)k for allk∈N, where|sr|> . Then we obtainBr,se=eand

Br,sv= (, , , . . .). Hence,eis in

∞but not inbr,sandvis inbr,sbut not in∞. This shows that the sequence spacesbr,sandoverlap but these spaces do not include each other.

This completes the proof.

Definition .(see []) An infinite matrixA= (ank) is called coregular, if A= (ank) is

conservative andχ(A) =limn→∞kankklimn→∞ank= .

By taking into account the regularity of the binomial matrix Br,s= (brnk,s), we obtain

χ(Br,s) = = . So, the binomial matrixBr,s= (br,s

nk) is coregular.

Definition .(see []) LetA= (ank) be an infinite matrix with bounded columns. Then

Ais defined to be of typeMiftA=  impliest=  for everyt.

Definition .(see []) For a conservative triangleA= (ank),ccA. Its closurec¯incAis

called the perfect part ofcA. Ifcis dense,A= (ank) is called perfect.

Now we give the following two theorems, which are needed.

Theorem .(see []) A regular triangle A= (ank)is of type M if there exists a(zi)∈

with zi=zj(i=j)and(zi)∞< such that

∀i∈N,∃(xki)k∞, ∀n∈N:z

n i =

n

k= ankxki.

Theorem .(see []) A coregular triangle is perfect if and only if it is of type M.

Theorem . Each regular binomial matrix Br,s= (brnk,s)is perfect. Proof We know that the regular binomial matrixBr,s= (br,s

nk) is coregular. So, for the proof,

we should show thatBr,s= (br,s

nk) is of typeM.

LetDr,s= (dr,s

nk) be the inverse ofBr,s= (b r,s

nk). Then we have for everyz∈C

uk(z) = k

v=

drkv,szv=  rk

k

v=

k v

(–s)kv(s+r)vzv

=  rk

(s+r)zsk.

One can easily verify thatsupkNsup{|uk(z)|:|(s+r)zs|<|r|} ≤ and that n

k=

brnk,suk(z) = n

k= brnk,s

k

v=

(6)

for all z∈C. Therefore, if we choose a sequence (zi) withzi=zj(i=j),(zi)∞<  and

|(s+r)zis|<|r|(i∈N), thenxki=uk(zi) (i,k∈N) fit Theorem .. Thus, the regular binomial matrixBr,s= (brnk,s) is perfect. This completes the proof.

3 The Schauder basis and

α

-,

β

-,

γ

- and continuous duals

In this chapter, we construct the Schauder basis and designate theα-,β-,γ-, and contin-uous duals of the binomial sequence spacesbr,sandbr,s

c .

Given a normed space (X, · X), a set{xk:xkX,k∈N}is called a Schauder basis for

Xif for allxXthere exist unique scalarsλk,k∈N, such thatx=

kλkxk;i.e.,

x

n

k=

λkxk

X

−→

asn→ ∞[].

Theorem . Let μk={Br,sx}k for all k∈Nand l=limk→∞μk. We define a sequence

g(k)(r,s) ={g(k)

n (r,s)}n∈Nas follows:

gn(k)(r,s) = ⎧ ⎨ ⎩

, ≤n<k,

rn n

k

(–s)nk(s+r)k, nk,

for all fixed k∈N.Then the following statements hold. (a) The sequence{g(k)(r,s)}

k∈Nis a Schauder basis for the binomial sequence spacebr,s,

and everyxbr,shas a unique representation of the form

x=

k

μkg(k)(r,s).

(b) The set{e,g()(r,s),g()(r,s), . . .}is a Schauder basis for the binomial sequence space br,s

c ,and anyxbrc,shas a unique representation of the form

x=le+

k

[μkl]g(k)(r,s).

Proof (a) Obviously we have

Br,sg(k)(r,s) =e(k)∈c, k∈N,

where e(k) is a sequence with  in thekth place and zeros elsewhere. So, the inclusion {g(k)(r,s)} ⊂br,s

 holds.

Given a sequencex= (xk)∈br,sandm∈N, we define

x[m]=

m

k=

μkg(k)(r,s).

Then, if we apply the binomial matrixBr,s= (br,s

nk) tox[m], we have

Br,sx[m]=

m

k=

μkBr,sg(k)(r,s) = m

k=

(7)

and

Br,sxx[m]n= ⎧ ⎨ ⎩

, ≤nm,

(Br,sx)

n, n>m,

for allm,n∈N.

Now, let >  be arbitrarily given. Then we may choose am∈Nsuch that

Br,sxm< 

for allmm. Therefore,

xx[m]br,s

 =supnm

Br,sxn< sup

nm

Br,sxn≤ <

for allmm. This shows that

x=

k

μkg(k)(r,s).

To complete the proof of part (a), we should show the uniqueness of this representation. We assume that there exists a representation

x=

k

λkg(k)(r,s).

Due to the transformation,Ldefined in the proof of Theorem . is continuous; we can write

Br,sxn=

k

λk

Br,sg(k)(r,s)n=

k

λke(nk)=λn

for alln∈N, which contradicts the fact that (Br,sx)

n=μnfor alln∈N. Hence, everyxbr,s has a unique representation, as desired.

(b) We know that{g(k)(r,s)} ⊂br,sandBr,se=ec. So, the inclusion{e,g(k)(r,s)} ⊂brc,s trivially holds.

For a given arbitrary sequencex= (xk)∈brc,s, we define a sequencey= (yk) such that

y=xle, wherel=limk→∞μk. Then it is obvious thaty= (yk)∈br,s. By considering the part (a), one can say thaty= (yk) has a unique representation. This implies thatx= (xk)

has a unique representation, as desired in part (b). This completes the proof.

By taking into account the results of Theorems . and ., we give the following result.

Corollary . The binomial sequence spaces br,sand br,s

c are separable.

LetXandYbe two arbitrary sequence spaces. The multiplier space ofXandYis sym-bolized withM(X,Y) and defined by

M(X,Y) =y= (yk)∈w:xy= (xkyk)∈Yfor allx= (xk)∈X

(8)

By considering the definition ofM(X,Y), theα-,β-, andγ-duals of a sequence space X are defined by

=M(X,), =M(X,cs) and =M(X,bs),

respectively.

ByX∗, we denote the space of all bounded linear functionals onX.X∗is called the con-tinuous dual of a normed spaceX.

Let us give some properties to use in the next lemma:

sup

KF

n

kK

ank

p<∞, (.)

sup

n∈N

k

|ank|<∞, (.)

lim

n→∞ank=αk for eachk∈N, (.)

lim

n→∞

k

ank=α, (.)

whereFis the collection of all finite subsets ofNand ≤p<∞.

Lemma .(see []) Let A= (ank)be an infinite matrix.Then the following statements

hold:

(i) A= (ank)∈(c:) = (c:)⇔(.)holds withp= ;

(ii) A= (ank)∈(c:c)⇔(.)and(.)hold;

(iii) A= (ank)∈(c:c)⇔(.), (.)and(.)hold;

(iv) A= (ank)∈(c:∞) = (c:∞)⇔(.)holds;

(v) A= (ank)∈(c:p)⇔(.)holds with≤p<∞.

Theorem . Theα-dual of the binomial sequence spaces br,sand br,s c is

vr,s=

a= (ak)∈w:sup KF

n

kK

n k

(–s)nkrn(r+s)kan

<∞

.

Proof Let us consider the sequencex= (xn) that is defined in the proof of Theorem ..

Then, for givena= (an)∈w, we obtain

anxn= n

k=

n k

(–s)nkrn(r+s)kanyk=

Ur,syn

for alln∈N. By considering the equality above, we deduce thatax= (anxn)∈whenever x= (xk)∈br,sorbrc,sif and only ifUr,sywhenevery= (yk)∈corc. This shows that a= (an)∈ {br,s}α={bcr,s}α if and only if Ur,s∈(c:) = (c:). If we combine this and Lemma .(i), we obtain

a= (an)∈

br,=brc, ⇔ sup

KF

n

kK

n k

(–s)nkrn(r+s)kan

<∞.

Therefore,{br,s}α={br,s c }α=v

r,s

(9)

Theorem . Define the sets vr,s,vr,s,and vr,sby

vr,s=

a= (ak)∈w:sup n∈N

n k= n

j=k

j k

(–s)jkrj(r+s)kaj

<∞

,

vr,s=

a= (ak)∈w:

j=k

j k

(–s)jkrj(r+s)kajexists for each k∈N

,

vr,s=

a= (ak)∈w: lim n→∞ n k= n

j=k

j k

(–s)jkrj(r+s)kajexists

.

Then the following statements hold. (I) {br,s}β=vr,s

 ∩vr,s, (II) {br,s

c }β=vr,svr,svr,s, (III) {br,s}γ ={br,s

c }γ=v r,s

 .

Proof Leta= (an)∈w. If we bear in mind the sequencex= (xk) that is defined in proof of

Theorem ., we have

n

k= akxk=

n k=  rk k j= k j

(–s)kj(r+s)jyj

ak = n k= n

j=k

j k

(–s)jkrj(r+s)kaj

yk

=Hr,syn

for alln∈N, where the matrixHr,s= (hr,s

nk) is defined by

hrnk,s= ⎧ ⎨ ⎩

n j=k

j k

(–s)jkrj(r+s)ka

j, ≤kn,

, k>n,

for allk,n∈N. Then:

(I)ax= (akxk)∈cswheneverx= (xk)∈br,sif and only ifHr,sycwhenevery= (yk)∈

c. This shows thata= (ak)∈ {br,s}β if and only ifHr,s∈(c:c). If we combine this and Lemma .(ii), we conclude that

sup

n∈N

n

k=

hrnk,s<∞, (.)

lim

n→∞h

r,s

nk exists for eachk∈N. (.)

These results show that{br,s}β=vr,s

 ∩v

r,s

 .

(II) In a similar way, we obtaina= (ak)∈ {brc,s}βif and only ifHr,s∈(c:c). If we combine

this and Lemma .(iii), we deduce that (.), (.) hold and

lim

n→∞ n

k=

hrnk,s exists, (.)

which shows that{br,s c }β=v

r,s

 ∩v

r,s

 ∩v

r,s

(10)

(III)ax= (akxk)∈bswheneverx= (xk)∈br,sorbrc,sif and only ifHr,sy∞ whenever

y= (yk)∈corc. This means thata= (ak)∈ {br,s}γ={brc,s}γ if and only ifHr,s∈(c:∞) = (c:∞). By combining this and Lemma .(iv), we deduce that (.) holds. Hence,{br,s}γ= {br,s

c }γ=vr,s. This completes the proof.

Theorem . {br,s}∗and{br,s

c }∗are equivalent to.

Proof To avoid a repetition of similar statements, the proof of the theorem is given for only the binomial sequence spacebr,s

c . For the proof, the existence of a linear surjective

norm preserving mappingL:{brc,s}∗−→should be shown. Suppose thatf∈ {br,s

c }∗. Now from Theorem .(b) we know that{e,g()(r,s),g()(r,s), . . .}

is a basis forbr,s

c , and eachxbrc,shas a unique representation of the form

x=le+

k

[μkl]g(k)(r,s).

By the linearity and continuity off, we get

f(x) =lf(e) +

k

[μkl]f

g(k)(r,s) (.)

for allxbrc,s. Now define a sequencex= (xk)∈brc,ssuch thatxbrc,s=  as follows:

xk=

⎧ ⎨ ⎩ 

rk k

j= k

j

(–s)kj(r+s)jsgnf(g(j)(r,s)), kn,

rk n

j= k

j

(–s)kj(r+s)jsgnf(g(j)(r,s)), k>n,

for allk,n∈N. Hence,

f(x)=

n

k=

fg(k)(r,s)f (.)

since|f(x)| ≤ f · xonbr,s

c . It follows from (.) that

k

fg(k)(r,s)=sup

n∈N

n

k=

fg(k)(r,s)f.

Now we write (.) as

f(x) =al+

k

akμk,

where

a=f(e) –

k

fg(k)(r,s), ak=f

g(k)(r,s)

(11)

By taking into account|limk→∞(Br,sx)k| ≤ xbrc,s= , we get f(x)≤ xbr,s

c

|a|+

k

|ak|

whence

f(x)≤ |a|+

k

|ak| (.)

and|f(x)| ≤ f, so we define for alln≥,

xk=

⎧ ⎨ ⎩ 

rk k

j= k

j

(–s)kj(r+s)jsgna

j, ≤kn,

rk n

j= k

j

(–s)kj(r+s)jsgna j+rk

k j=n+

k j

(–s)kj(r+s)jsgna, k>n.

Thenxbr,s

c ,xbrc,s= ,limk→∞(B

r,sx)

k=sgnaand so

f(x)= |a|+

n

k=

|ak|+sgna

k=n+ ak

≤ f.

We know thatlimn→∞k=n+ak=  whenevera= (ak)∈. Then, if we pass to the limit asn→ ∞in the last inequality, we have

|a|+

k

|ak| ≤ f. (.)

By combining (.) and (.), we conclude that

f=|a|+

k

|ak|,

which is the norm on.

Let us define a transformationLsuch thatL:{br,s

c }∗−→,L(f) = (a,a,a, . . .). Then we

have

L(f)=|a|+|a|+|a|+· · ·=f

L(f)being thenorm. Therefore,Lis norm preserving. It is obvious thatLis surjective

and linear. This completes the proof.

4 Some matrix classes related to the binomial sequence spaces

In this chapter, we characterize some matrix classes related to the binomial sequence spaces br,sandbr,s

c . Now, we start with two lemmas which are required in the proof of

the *theorems.

Lemma .(see []) Each matrix map between BK -spaces is continuous.

(12)

For brevity of notations, here and in the following, we prefer to use

tnkr,s=

n

j=k

j k

(–s)jkrj(r+s)kanj

for alln,k∈N.

Theorem . Let A= (ank)be an infinite matrix with complex entries.Then the following

statements hold.

(I) A∈(brc,s:p)if and only if

sup

KF

n

kK

tnkr,s

p

<∞, (.)

tnkr,s exists for allk,n∈N, (.)

k

trnk,s converges for alln∈N, (.)

sup

m∈N

m

k=

m

j=k

j k

(–s)jkrj(r+s)kanj

<∞, n∈N, (.)

where≤p<∞. (II) A∈(br,s

c :∞)if and only if (.)and(.)hold,and

sup

n∈N

k

trnk,s<∞. (.)

Proof Given a sequencex= (xk)∈brc,s, we suppose that the conditions (.)-(.) hold.

Then, by taking into account Theorem .(II), we conclude that{ank}k∈N∈ {brc,s}β for all

n∈N. Thus, theA-transform ofxexists. Let us consider a matrixUr,s= (ur,s

nk) defined by

urnk,s=trnk,sfor alln,k∈N. SinceUr,s= (ur,s

nk) satisfies Lemma .(v), we deduce thatUr,s=

(urnk,s)∈(c:p).

Now, we consider the following equality:

m

k= ankxk=

m

k=

m

j=k

j k

(–s)jkrj(r+s)kanjyk (.)

for alln,m∈N. If we take the limit (.) side by side asm→ ∞, we obtain

k

ankxk=

k

trnk,syk. (.)

By taking thep-norm (.) side by side, we have

Axp=U

r,sy

p<∞.

(13)

Conversely, suppose thatA∈(br,s

c :p). It is known thatbrc,sandpareBK-spaces. If we

combine this fact and Lemma ., we deduce that there is a constantM>  such that

Axp≤Mxbrc,s (.)

for allxbr,s c .

Now, we define a sequencex= (xk) such thatx=kKg(k)(r,s) for all fixedk∈N, where

g(k)(r,s) ={g(nk)(r,s)}n∈NandKF. Since inequality (.) holds for allxbrc,s, we have

Axp=

n

kK

trnk,s

pp

Mxbrc,s=M.

Therefore (.) holds.

According to the assumption,Acan be applied to the binomial sequence spacebr,s c . So,

it is trivial that the conditions (.)-(.) hold. This completes the proof of part (I). If we take Lemma .(iv) instead of Lemma .(v), then part (II) can be proved in a

similar way.

Theorem . Let A= (ank)be an infinite matrix with complex entries.Then,A∈(brc,s:c)

if and only if (.), (.),and(.)hold,and

lim

n→∞t r,s

nk=αk for all k∈N, (.)

lim

n→∞

k

tnkr,s=α. (.)

Proof Suppose thatAsatisfies the conditions (.), (.), (.), (.), and (.). Given an arbitrary sequencex= (xk)∈brc,swithlimk→∞xk=l, thenAxexists. SinceBr,s= (brnk,s) is

regular andy= (yk) is connected with the sequencex= (xk) by equation (.), we obtain

y= (yk)∈csuch thatlimk→∞yk=l.

By considering the conditions (.) and (.), we have

k

j=

|αj| ≤sup n∈N

j

tnjr,s<∞

for allk∈N. This shows us that (αk)∈. Bearing in mind the condition (.), we have

k

ankxk=

k

trnk,s(ykl) +l

k

trnk,s, n∈N. (.)

By taking into account the conditions (.), (.), and (.), if we take the limit (.) side by side asn→ ∞, we write

lim

n→∞(Ax)n=

k

αk(ykl) +lα, (.)

(14)

On the contrary, suppose thatA∈(br,s

c :c). It is well known that every convergent

se-quence is also bounded, namelyc∞. By combining this fact and Theorem .(II), we

deduce that the necessity of the conditions (.), (.), and (.) holds. SinceAxexists and belongs tocfor allxbrc,s, if we takeg(k)(r,s) ={gn(k)(r,s)}n∈N instead of an arbitrary

sequencex= (xk), we deduce thatAg(k)(r,s) ={tnkr,s}n∈N∈cfor allk∈N. This shows us that

the necessity of (.) holds.

Moreover, if we takex=ein (.), we obtainAx={ktnkr,s}n∈N∈c. The last result is the

necessity of (.). This completes the proof.

Corollary . Let A= (ank)be an infinite matrix with complex entries.Then A∈(brc,s:c)m

if and only if the conditions(.), (.),and(.)hold,and the conditions(.)and(.) hold withαk= for all k∈Nandα=m,in turn.

Lemma .(see []) Let A= (ank)be an infinite matrix with complex entries.Then A

(br,s

∞:c)if and only if (.)and(.)hold,and

lim

n→∞

k

tr,s nk=

k

lim

n→∞t

r,s

nk, (.)

lim

m→∞

k

m

j=k

j k

(–s)jkrj(r+s)kanj

=

k

tnkr,s, n∈N. (.)

Theorem . (brc,s:c)m∩(br∞,s:c) =∅.

Proof We suppose that (br,s

c :c)m∩(b∞r,s:c)=∅. Then there is at least a matrixA= (ank)

such that the conditions of Corollary . and Lemma . hold forA= (ank). If we consider

the conditions (.) and (.), we conclude that

lim

n→∞

k

trnk,s= .

This result contradicts the condition (.). So, the classes (brc,s:c)mand (br∞,s:c) are

dis-joint. This last step completes the proof of the theorem.

Now, by using Lemma ., we can give some more results.

Corollary . Given an infinite matrix A= (ank)with complex entries,we define a matrix

Eu,v= (eu,v

nk)as follows:

eunk,v=  (u+v)n

n

j=

n j

vnjujajk

for all n,k∈N,where u,v∈Rand uv> .Then the necessary and sufficient conditions in order that A belongs to any of the classes(br,s

c :bu∞,v), (bcr,s:bup,v), (bcr,s:buc,v)and(bcr,s:buc,v)m

are obtained by taking Eu,v= (eu,v

nk)instead of A= (ank)in the required ones in Theorems

(15)

Corollary . Given an infinite matrix A= (ank)with complex entries,we define a matrix

C= (cnk)as follows:

cnk=

n+ 

n

j= ajk

for all n,k∈N.Then the necessary and sufficient conditions in order that A belongs to any of the classes(br,s

c :X∞), (bcr,s:Xp), (brc,s:c),˜ and(brc,s:c)˜mare obtained by taking C= (cnk)

instead of A= (ank)in the required ones in Theorems., .and Corollary.,where Xp,

X∞,andc are defined in˜ []and[],respectively.

Corollary . Given an infinite matrix A= (ank)with complex entries, we define two

matrices C= (cnk)and E= (enk)as follows:

cnk=ankan+,k and enk=ankan–,k

for all n,k∈N.Then the necessary and sufficient conditions in order that A belongs to any of the classes(br,s

c :∞()), (brc,s:c()), (bcr,s:p()),and(brc,s:c())mare obtained by taking

C= (cnk)or E= (enk)instead of A= (ank)in the required ones in Theorems., .and

Corollary.,where∞()and c()are defined in[]andp()is defined in[].

Corollary . Given an infinite matrix A= (ank)with complex entries,we define a matrix

E= (enk)as follows:

enk=

n+ 

n

j=

 +tja jk

for all n,k∈N,where <t< .Then the necessary and sufficient conditions in order that A belongs to any of the classes(br,s

c :at∞), (brc,s:atp), (bcr,s:atc),and(bcr,s:atc)mare obtained by

taking E= (enk)instead of A= (ank)in the required ones in Theorems., .and

Corol-lary.,where at

∞,atp,and atcare defined in[]and[],respectively.

Corollary . Given an infinite matrix A= (ank)with complex entries,we define a matrix

E= (enk)as follows:

enk= n

j= ajk

for all n,k∈N.Then the necessary and sufficient conditions in order that A belongs to any of the classes(br,s

c :bs), (brc,s:cs),and(bcr,s:cs)mare obtained by taking E= (enk)instead of

A= (ank)in the required ones in Theorems., .and Corollary..

5 Conclusion

By considering the definition of the binomial matrixBr,s= (br,s

nk), we deduce thatBr,s= (b r,s nk)

(16)

means of orderr. Moreover, the binomial matrixBr,s= (br,s

nk) is not a special case of the

weighed mean matrices. So, this paper filled up a gap in the existent literature.

Competing interests

The authors declare that they have no competing interests.

Author’s contributions

The authors read and approved the final manuscript.

Acknowledgements

We would like to express our thanks to the anonymous reviewers for their valuable comments.

Received: 13 August 2016 Accepted: 22 November 2016

References

1. Choudhary, B, Nanda, S: Functional Analysis with Applications. Wiley, New Delhi (1989)

2. Wilansky, A: Summability Through Functional Analysis. North-Holland Mathematics Studies, vol. 85. Elsevier, Amsterdam (1984)

3. Wang, C-S: On Nörlund sequence spaces. Tamkang J. Math.9, 269-274 (1978)

4. Ng, P-N, Lee, P-Y: Cesàro sequence spaces of non-absolute type. Comment. Math. (Prace Mat.)20(2), 429-433 (1978) 5. ¸Sengönül, M, Ba¸sar, F: Some new Cesàro sequence spaces of non-absolute type which include the spacesc0andc.

Soochow J. Math.31(1), 107-119 (2005)

6. Aydın, C, Ba¸sar, F: On the new sequence spaces which include the spacesc0andc. Hokkaido Math. J.33(2), 383-398

(2004)

7. Altay, B, Ba¸sar, F: Some Euler sequence spaces of non-absolute type. Ukr. Math. J.57(1), 1-17 (2005)

8. Altay, B, Ba¸sar, F, Mursaleen, M: On the Euler sequence spaces which include the spacespand∞I. Inf. Sci.176(10), 1450-1462 (2006)

9. Mursaleen, M, Ba¸sar, F, Altay, B: On the Euler sequence spaces which include the spacespand∞II. Nonlinear Anal. 65(3), 707-717 (2006)

10. Altay, B, Polat, H: On some new Euler difference sequence spaces. Southeast Asian Bull. Math.30(2), 209-220 (2006) 11. Polat, H, Ba¸sar, F: Some Euler spaces of difference sequences of order m. Acta Math. Sci. Ser. B, Engl. Ed.27B(2),

254-266 (2007)

12. Kara, EE, Ba¸sarır, M: On compact operators and some EulerB(m)-difference sequence spaces. J. Math. Anal. Appl.

379(2), 499-511 (2011)

13. Karakaya, V, Polat, H: Some new paranormed sequence spaces defined by Euler difference operators. Acta Sci. Math. (Szeged)76, 87-100 (2010)

14. Demiriz, S, Çakan, C: On some new paranormed Euler sequence spaces and Euler core. Acta Math. Sin. Engl. Ser. 26(7), 1207-1222 (2010)

15. Boos, B: Classical and Modern Methods in Summability. Oxford University Press, New York (2000)

16. Stieglitz, M, Tietz, H: Matrix Transformationen von Folgenräumen eine Ergebnisübersicht. Math. Z.154, 1-16 (1977) 17. Ba¸sar, F, Altay, B: On the space of sequences of p-bounded variation and related matrix mappings. Ukr. Math. J.55(1),

136-147 (2003)

18. Bisgin, MC: Matrix Transformations And Compact Operators On The Binomial Sequence Spaces. Under review 19. Kızmaz, H: On certain sequence spaces. Can. Math. Bull.24(2), 169-176 (1981)

References

Related documents