Some topological and geometrical properties of new Banach sequence spaces

In the present paper, we introduce a new band matrix \hat{F} and define the sequence space

where f_k is the k th Fibonacci number for every k\in \mathbb{N}. We also establish some inclusion relations concerning this space and determine its α-, β-, γ-duals. Further, we characterize some matrix classes on the space {\ell }_p(\hat{F}) and examine some geometric properties of this space.

MSC:11B39, 46A45, 46B45, 46B20.

Let ω be the space of all real-valued sequences. Any vector subspace of ω is called a sequence space. By {\ell }_{\mathrm{\infty }}, c, c_0 and {\ell }_p (), we denote the sets of all bounded, convergent, null sequences and p-absolutely convergent series, respectively. Also, we use the conventions that e=(1,1,\dots ) and e^{(n)} is the sequence whose only non-zero term is 1 in the n th place for each n\in \mathbb{N}, where \mathbb{N}=\{0,1,2,\dots \}.

Let X and Y be two sequence spaces and A=(a_{nk}) be an infinite matrix of real numbers a_{nk}, where n,k\in \mathbb{N}. We write A=(a_{nk}) instead of A={(a_{nk})}_{n,k=0}^{\mathrm{\infty }}. Then we say that A defines a matrix mapping from X into Y and we denote it by writing A:X\to Y if for every sequence x={(x_k)}_{k=0}^{\mathrm{\infty }}\in X, the sequence Ax={\{A_n(x)\}}_{n=0}^{\mathrm{\infty }}, the A-transform of x, is in Y, where

For simplicity in notation, here and in what follows, the summation without limits runs from 0 to ∞. Also, if x\in \omega, then we write x=(x_k) instead of x={(x_k)}_{k=0}^{\mathrm{\infty }}.

By (X,Y), we denote the class of all matrices A such that A:X\to Y. Thus, A\in (X,Y) if and only if the series on the right-hand side of (1.1) converges for each n\in \mathbb{N} and every x\in X and we have Ax\in Y for all x\in X.

The matrix domain X_A of an infinite matrix A in a sequence space X is defined by

which is a sequence space.

The approach constructing a new sequence space by means of the matrix domain of a particular limitation method has recently been employed by several authors; see, for instance, [1–12].

Let Δ denote the matrix \mathrm{\Delta }=({\mathrm{\Delta }}_{nk}) defined by

or

In the literature, the matrix domain {\lambda }_{\mathrm{\Delta }} is called the difference sequence space whenever λ is a normed or paranormed sequence space. The idea of difference sequence spaces was introduced by Kızmaz [13]. In 1981, Kızmaz [13] defined the sequence spaces

for X={\ell }_{\mathrm{\infty }}, c and c_0. The difference space b{v}_p, consisting of all sequences (x_k) such that (x_k-x_{k-1}) is in the sequence space {\ell }_p, was studied in the case by Altay and Başar [14] and in the case 1\le p\le \mathrm{\infty } by Başar and Altay [4] and Çolak et al. [15]. The paranormed difference sequence space

was examined by Ahmad and Mursaleen [16] and Malkowsky [17], where \lambda (p) is any of the paranormed spaces {\ell }_{\mathrm{\infty }}(p), c(p) and c_0(p) defined by Simons [18] and Maddox [19].

Recently, Başar et al. [20] have defined the sequence spaces bv(u,p) and b{v}_{\mathrm{\infty }}(u,p) by

and

where u=(u_k) is an arbitrary fixed sequence and for all k\in \mathbb{N}. These spaces are generalization of the space b{v}_p for 1\le p\le \mathrm{\infty }. Quite recently, Kirişçi and Başar [21] have introduced and studied the generalized difference sequence spaces

for X={\ell }_{\mathrm{\infty }}, {\ell }_p, c and c_0, where and B(r,s)x=(s{x}_{k-1}+r{x}_k) (r,s\ne 0). Following Kirişçi and Başar [21], Sönmez [22] has examined the sequence space X(B) as the set of all sequences whose B(r,s,t)-transforms are in the space X\in \{{\ell }_{\mathrm{\infty }},{\ell }_p,c,c_0\}, where B(r,s,t) denotes the triple band matrix B(r,s,t)=\{b_{nk}(r,s,t)\} defined by

for all n,k\in \mathbb{N} and r,s,t\in \mathbb{R}-\{0\}. Also in [23–34], the authors studied some difference sequence spaces.

In this paper, we define the Fibonacci difference matrix \hat{F} by using the Fibonacci sequence {\{f_n\}}_{n=0}^{\mathrm{\infty }} and introduce new sequence spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}) related to the matrix domain of \hat{F} in the sequence spaces {\ell }_p and {\ell }_{\mathrm{\infty }}, respectively, where . This study is organized as follows.

In Section 2, we give some notations and basic concepts including the Fibonacci sequence and a BK-space. In Section 3, we define a new band matrix with Fibonacci numbers and introduce the sequence spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}). Also, we establish some inclusion relations concerning these spaces and construct the basis of the space {\ell }_p(\hat{F}) for . In Section 4, we determine the α-, β-, γ-duals of the spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}). In Section 5, we characterize the classes ({\ell }_p(\hat{F}),X) and ({\ell }_{\mathrm{\infty }}(\hat{F}),X), where and X is any of the spaces {\ell }_{\mathrm{\infty }}, {\ell }_1, c and c_0. In the final section of the paper, we investigate some geometric properties of the space {\ell }_p(\hat{F}) for .

Define the sequence {\{f_n\}}_{n=0}^{\mathrm{\infty }} of Fibonacci numbers given by the linear recurrence relations

Fibonacci numbers have many interesting properties and applications in arts, sciences and architecture. For example, the ratio sequences of Fibonacci numbers converges to the golden ratio which is important in sciences and arts. Also, some basic properties of Fibonacci numbers [35] are given as follows:

Substituting for f_{n+1} in Cassini’s formula yields f_{n-1}^2+f_n{f}_{n-1}-f_n^2={(-1)}^{n+1}.

A sequence space X is called a FK-space if it is a complete linear metric space with continuous coordinates p_n:X\to \mathbb{R} (n\in \mathbb{N}), where ℝ denotes the real field and p_n(x)=x_n for all x=(x_k)\in X and every n\in \mathbb{N}. A BK space is a normed FK space, that is, a BK-space is a Banach space with continuous coordinates. The space {\ell }_p () is a BK-space with {\parallel x\parallel }_p={({\sum }_{k=0}^{\mathrm{\infty }}|x_k{|}^p)}^{1/p} and c_0, c and {\ell }_{\mathrm{\infty }} are BK-spaces with {\parallel x\parallel }_{\mathrm{\infty }}={sup}_k|x_k|.

A sequence (b_n) in a normed space X is called a Schauder basis for X if for every x\in X, there is a unique sequence ({\alpha }_n) of scalars such that x={\sum }_n{\alpha }_n{b}_n, i.e., {lim}_{m\to \mathrm{\infty }}\parallel x-{\sum }_{n=0}^m{\alpha }_n{b}_n\parallel =0.

The α-, β- and γ-duals of the sequence space X are respectively defined by

and

where cs and bs are the sequence spaces of all convergent and bounded series, respectively [36].

We assume throughout that p,q\ge 1 with p^{-1}+q^{-1}=1 and denote the collection of all finite subsets of ℕ by ℱ.

In this section, we define the Fibonacci band matrix \hat{F}=({\hat{f}}_{nk}) and introduce the sequence spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}), where . Also, we present some inclusion theorems and construct the Schauder basis of the space {\ell }_p(\hat{F}) for.

Let f_n be the n th Fibonacci number for every n\in \mathbb{N}. Then we define the infinite matrix \hat{F}=({\hat{f}}_{nk}) by

Now, we introduce the Fibonacci difference sequence spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}) as the set of all sequences such that their \hat{F}-transforms are in the space {\ell }_p and {\ell }_{\mathrm{\infty }}, respectively, i.e.,

and

With the notation of (1.2), the sequence spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}) may be redefined by

Define the sequence y=(y_n), which will be frequently used, by the \hat{F}-transform of a sequence x=(x_n), i.e.,

Now, we may begin with the following theorem which is essential in the text.

Theorem 3.1 Let 1\le p\le \mathrm{\infty }. Then {\ell }_p(\hat{F}) is a BK-space with the norm , that is,

and

Proof Since (3.1) holds, {\ell }_p and {\ell }_{\mathrm{\infty }} are BK-spaces with respect to their natural norms and the matrix \hat{F} is a triangle; Theorem 4.3.12 of Wilansky [[37], p.63] gives the fact that the spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}) are BK-spaces with the given norms, where . This completes the proof. □

Remark 3.2 One can easily check that the absolute property does not hold on the spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}), that is, {\parallel x\parallel }_{{\ell }_p(\hat{F})}\ne {\parallel |x|\parallel }_{{\ell }_p(\hat{F})} and {\parallel x\parallel }_{{\ell }_{\mathrm{\infty }}(\hat{F})}\ne {\parallel |x|\parallel }_{{\ell }_{\mathrm{\infty }}(\hat{F})} for at least one sequence in the spaces {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}), and this shows that {\ell }_p(\hat{F}) and {\ell }_{\mathrm{\infty }}(\hat{F}) are the sequence spaces of non-absolute type, where |x|=(|x_k|) and .

Theorem 3.3 The Fibonacci difference sequence space {\ell }_p(\hat{F}) of non-absolute type is linearly isomorphic to the space {\ell }_p, that is, {\ell }_p(\hat{F})\cong {\ell }_p for 1\le p\le \mathrm{\infty }.

Proof To prove this, we should show the existence of a linear bijection between the spaces {\ell }_p(\hat{F}) and {\ell }_p for 1\le p\le \mathrm{\infty }. Consider the transformation T defined, with the notation of (3.2), from {\ell }_p(\hat{F}) to {\ell }_p by x\to y=Tx. Then Tx=y=\hat{F}x\in {\ell }_p for every x\in {\ell }_p(\hat{F}). Also, the linearity of T is clear. Further, it is trivial that x=0 whenever Tx=0 and hence T is injective.

Furthermore, let y=(y_k)\in {\ell }_p for 1\le p\le \mathrm{\infty } and define the sequence x=(x_k) by

Then, in the cases and p=\mathrm{\infty }, we get

and

respectively. Thus, we have x\in {\ell }_p(\hat{F}) (1\le p\le \mathrm{\infty }). Hence, T is surjective and norm preserving. Consequently, T is a linear bijection which shows that the spaces {\ell }_p(\hat{F}) and {\ell }_p are linearly isomorphic for 1\le p\le \mathrm{\infty }. This concludes the proof. □

Now, we give some inclusion relations concerning the space {\ell }_p(\hat{F}).

Theorem 3.4 The inclusion {\ell }_p\subset {\ell }_p(\hat{F}) strictly holds for 1\le p\le \mathrm{\infty }.

Proof To prove the validity of the inclusion {\ell }_p\subset {\ell }_p(\hat{F}) for 1\le p\le \mathrm{\infty }, it suffices to show the existence of a number M>0 such that {\parallel x\parallel }_{{\ell }_p(\hat{F})}\le M{\parallel x\parallel }_p for every x\in {\ell }_p.

Let x\in {\ell }_p and . Since the inequalities \frac{f_k}{f_{k+1}}\le 1 and \frac{f_{k+1}}{f_k}\le 2 hold for every k\in \mathbb{N}, we obtain with the notation of (3.2),

and

which together yield, as expected,

(3.4)

for . Further, since the sequence x=(x_k)=(f_{k+1}^2)=(1,2^2,3^2,5^2,\dots ) is in {\ell }_p(\hat{F})-{\ell }_p, the inclusion {\ell }_p\subset {\ell }_p(\hat{F}) is strict for . Similarly, one can easily prove that inequality (3.4) also holds in the case p=1, and so we omit the details. This completes the proof. □

Theorem 3.5 Neither of the spaces b{v}_p and {\ell }_p(\hat{F}) includes the other one, where .

Proof Let e=(1,1,1,\dots ) and x=(x_k)=(f_{k+1}^2). Then, since \hat{F}x=(1,0,0,\dots )\in {\ell }_p and \mathrm{\Delta }x=(1,f_0{f}_3,f_1{f}_4,\dots ,f_{k-1}{f}_{k+2},\dots )\notin {\ell }_p, we conclude that x is in {\ell }_p(\hat{F}) but not in b{v}_p. Now, consider the equation

Then |{(-1)}^k-f_k{f}_{k+1}|>f_k{f}_{k+1} whenever k is odd, which implies that the series {\sum }_k|\frac{f_k}{f_{k+1}}-\frac{f_{k+1}}{f_k}{|}^p is not convergent, where . Thus, \hat{F}e=(\frac{f_k}{f_{k+1}}-\frac{f_{k+1}}{f_k}) is not in {\ell }_p for . Additionally, since \mathrm{\Delta }e=(1,0,0,\dots ), the sequence e is in {\ell }_p. Hence, the sequence spaces {\ell }_p(\hat{F}) and b{v}_p overlap but neither contains the other, as asserted. □

Theorem 3.6 If , then {\ell }_p(\hat{F})\subset {\ell }_s(\hat{F}).

Proof Let and x\in {\ell }_p(\hat{F}). Then we obtain from Theorem 3.1 that y\in {\ell }_p, where y is the sequence given by (3.2). Thus, the well-known inclusion {\ell }_p\subset {\ell }_s yields y\in {\ell }_s. This means that x\in {\ell }_s(\hat{F}) and hence, the inclusion {\ell }_p(\hat{F})\subset {\ell }_s(\hat{F}) holds. This completes the proof. □

Now, we give a sequence of the points of the space {\ell }_p(\hat{F}) which forms a basis for the space {\ell }_p(\hat{F}) ().

Theorem 3.7 Let and define the sequence c^{(k)}\in {\ell }_p(\hat{F}) for every fixed k\in \mathbb{N} by

Then the sequence {(c^{(k)})}_{k=0}^{\mathrm{\infty }} is a basis for the space {\ell }_p(\hat{F}), and every x\in {\ell }_p(\hat{F}) has a unique representation of the form

Proof Let . Then it is obvious by (3.5) that \hat{F}(c^{(k)})=e^{(k)}\in {\ell }_p (k\in \mathbb{N}) and hence c^{(k)}\in {\ell }_p(\hat{F}) for all k\in \mathbb{N}.

Further, let x\in {\ell }_p(\hat{F}) be given. For every non-negative integer m, we put

Then we have that

and hence

Now, for any given \epsilon >0, there is a non-negative integer m_0 such that

Therefore, we have for every m\ge m_0 that

which shows that {lim}_{m\to \mathrm{\infty }}{\parallel x-x^{(m)}\parallel }_{{\ell }_p(\hat{F})}=0 and hence x is represented as in (3.6).

Finally, let us show the uniqueness of the representation (3.6) of x\in {\ell }_p(\hat{F}). For this, suppose that x={\sum }_k{\mu }_k(x)c^{(k)}. Since the linear transformation T defined from {\ell }_p(\hat{F}) to {\ell }_p in the proof of Theorem 3.3 is continuous, we have

Hence, the representation (3.6) of x\in {\ell }_p(\hat{F}) is unique. This concludes the proof. □

In this section, we determine the α-, β- and γ-duals of the sequence space {\ell }_p(\hat{F}) of non-absolute type. Since the case p=1 can be proved by analogy, we omit the proof of that case and consider only the case in the proof of Theorems 4.5 and 4.6, respectively.

The following known results [38] are fundamental for our investigation.

Lemma 4.1 A=(a_{nk})\in ({\ell }_p,{\ell }_1) if and only if

Lemma 4.2 A=(a_{nk})\in ({\ell }_p,c) if and only if

(4.1)

(4.2)

Lemma 4.3 A=(a_{nk})\in ({\ell }_{\mathrm{\infty }},c) if and only if (4.1) holds and

Lemma 4.4 A=(a_{nk})\in ({\ell }_p,{\ell }_{\mathrm{\infty }}) if and only if (4.2) holds with .

Theorem 4.5 The α-dual of the space {\ell }_p(\hat{F}) is the set

where .

Proof Let . For any fixed sequence a=(a_n)\in \omega, we define the matrix B=(b_{nk}) by

for all n,k\in \mathbb{N}. Also, for every x=(x_n)\in \omega, we put y=\hat{F}x. Then it follows by (3.2) that

Thus, we observe by (4.4) that ax=(a_n{x}_n)\in {\ell }_1 whenever x\in {\ell }_p(\hat{F}) if and only if By\in {\ell }_1 whenever y\in {\ell }_p. Therefore, we derive by using Lemma 4.1 that

which implies that {({\ell }_p(\hat{F}))}^{\alpha }={\hat{d}}_1. □

Theorem 4.6 Define the sets {\hat{d}}_2, {\hat{d}}_3 and {\hat{d}}_4 by

and

Then {({\ell }_p(\hat{F}))}^{\beta }={\hat{d}}_2\cap {\hat{d}}_3 and {({\ell }_{\mathrm{\infty }}(\hat{F}))}^{\beta }={\hat{d}}_2\cap {\hat{d}}_4, where .

Proof Let a=(a_k)\in \omega and consider the equality

where D=(d_{nk}) is defined by

Then we deduce from Lemma 4.2 with (3.2) that ax=(a_k{x}_k)\in cs whenever x=(x_k)\in {\ell }_p(\hat{F}) if and only if Dy\in c whenever y=(y_k)\in {\ell }_p. Thus, (a_k)\in {({\ell }_p(\hat{F}))}^{\beta } if and only if (a_k)\in {\hat{d}}_2 and (a_k)\in {\hat{d}}_3 by (4.1) and (4.2), respectively. Hence, {({\ell }_p(\hat{F}))}^{\beta }={\hat{d}}_2\cap {\hat{d}}_3. It is clear that one can also prove the case p=\mathrm{\infty } by the technique used in the proof of the case with Lemma 4.3 instead of Lemma 4.2. So, we leave the detailed proof to the reader. □

Theorem 4.7 {({\ell }_p(\hat{F}))}^{\gamma }={\hat{d}}_3, where .

Proof This result can be obtained from Lemma 4.4 by using (4.5). □

In this section, we characterize the classes ({\ell }_p(\hat{F}),X), where 1\le p\le \mathrm{\infty } and X is any of the spaces {\ell }_{\mathrm{\infty }}, {\ell }_1, c and c_0.

For simplicity in notation, we write

for all k,n\in \mathbb{N}.

The following lemma is essential for our results.

Lemma 5.1 (see [[21], Theorem 4.1])

Let λ be an FK-space, U be a triangle, V be its inverse and μ be an arbitrary subset of ω. Then we have A=(a_{nk})\in ({\lambda }_U,\mu ) if and only if

and

where

and c_{nk}={\sum }_{j=k}^{\mathrm{\infty }}{a}_{nj}{v}_{jk} for all k,m,n\in \mathbb{N}.

Now, we list the following conditions:

(5.1)

(5.2)

(5.3)

(5.4)

(5.5)

(5.6)

(5.7)

(5.8)

(5.9)

(5.10)

(5.11)

(5.12)

Then, by combining Lemma 5.1 with the results in [38], we immediately derive the following results.

Theorem 5.2

1. (a)

   A=(a_{nk})\in ({\ell }_1(\hat{F}),{\ell }_{\mathrm{\infty }}) if and only (5.2), (5.9) and (5.10) hold.

2. (b)

   A=(a_{nk})\in ({\ell }_1(\hat{F}),c) if and only if (5.2), (5.6), (5.9) and (5.10) hold.

3. (c)

   A=(a_{nk})\in ({\ell }_1(\hat{F}),c_0) if and only if (5.2), (5.6) with {\tilde{\alpha }}_k=0, (5.9) and (5.10) hold.

4. (d)

   A=(a_{nk})\in ({\ell }_1(\hat{F}),{\ell }_1) if and only (5.2), (5.10) and (5.11) hold.

Theorem 5.3 Let . Then we have

1. (a)

   A=(a_{nk})\in ({\ell }_p(\hat{F}),{\ell }_{\mathrm{\infty }}) if and only if (5.1), (5.2) and (5.4) hold.

2. (b)

   A=(a_{nk})\in ({\ell }_p(\hat{F}),c) if and only if (5.1), (5.2), (5.4) and (5.6) hold.

3. (c)

   A=(a_{nk})\in ({\ell }_p(\hat{F}),c_0) if and only if (5.1), (5.2), (5.4) and (5.6) with {\tilde{\alpha }}_k=0 hold.

4. (d)

   A=(a_{nk})\in ({\ell }_p(\hat{F}),{\ell }_1) if and only if (5.1), (5.2) and (5.5) hold.

Theorem 5.4

1. (a)

   A=(a_{nk})\in ({\ell }_{\mathrm{\infty }}(\hat{F}),{\ell }_{\mathrm{\infty }}) if and only (5.2), (5.3) and (5.4) with q=1 hold.

2. (b)

   A=(a_{nk})\in ({\ell }_{\mathrm{\infty }}(\hat{F}),c) if and only (5.2), (5.3), (5.6) and (5.7) hold.

3. (c)

   A=(a_{nk})\in ({\ell }_{\mathrm{\infty }}(\hat{F}),c_0) if and only (5.2), (5.3) and (5.8) hold.

4. (d)

   A=(a_{nk})\in ({\ell }_{\mathrm{\infty }}(\hat{F}),{\ell }_1) if and only (5.2), (5.3) and (5.12) hold.

In this section, we study some geometric properties of the space {\ell }_p(\hat{F}) for .

For these properties, we refer to [3, 39–47].

A Banach space X is said to have the Banach-Saks property if every bounded sequence (x_n) in X admits a subsequence (z_n) such that the sequence \{t_k(z)\} is convergent in the norm in X [40], where

A Banach space X is said to have the weak Banach-Saks property whenever, given any weakly null sequence (x_n)\subset X, there exists a subsequence (z_n) of (x_n) such that the sequence \{t_k(z)\} is strongly convergent to zero.

In [43], García-Falset introduces the following coefficient:

where B(X) denotes the unit ball of X.

Remark 6.1 A Banach space X with has the weak fixed point property [44].

Let . A Banach space is said to have the Banach-Saks type p or the property {(BS)}_p if every weakly null sequence (x_k) has a subsequence (x_{k_l}) such that for some C>0,

for all n\in \mathbb{N} ( see [45]).

Now, we may give the following results related to some geometric properties, mentioned above, of the space {\ell }_p(\hat{F}), where .

Theorem 6.2 Let . Then the space {\ell }_p(\hat{F}) has the Banach-Saks type p.

Proof Let ({\epsilon }_n) be a sequence of positive numbers for which \sum {\epsilon }_n\le 1/2, and also let (x_n) be a weakly null sequence in B({\ell }_p(\hat{F})). Set z_0=x_0=0 and z_1=x_{n_1}=x_1. Then there exists m_1\in \mathbb{N} such that

Since (x_n) being a weakly null sequence implies x_n\to 0 coordinatewise, there is an n_2\in \mathbb{N} such that

when n\ge n_2. Set z_2=x_{n_2}. Then there exists an m_2>m_1 such that

Again using the fact that x_n\to 0 coordinatewise, there exists an n_3\ge n_2 such that

when n\ge n_3.

If we continue this process, we can find two increasing subsequences (m_i) and (n_i) such that

for each n\ge n_{j+1} and

where b_j=x_{n_j}. Hence,

On the other hand, it can be seen that . Therefore, we have that

Hence, we obtain

By using the fact that 1\le {(n+1)}^{1/p} for all n\in \mathbb{N} and , we have

Hence, {\ell }_p(\hat{F}) has the Banach-Saks type p. This concludes the proof. □

Remark 6.3 Note that R({\ell }_p(\hat{F}))=R({\ell }_p)=2^{1/p} since {\ell }_p(\hat{F}) is linearly isomorphic to {\ell }_p.

Hence, by Remarks 6.1 and 6.3, we have the following theorem.

Theorem 6.4 The space {\ell }_p(\hat{F}) has the weak fixed point property, where .