The Banach contraction principle in $C^{*}$-algebra-valued b-metric spaces with application

Abstract

We introduce the notion of a $C^{*}$-algebra-valued b-metric space. We generalize the Banach contraction principle in this new setting. As an application of our result, we establish an existence result for an integral equation in a $C^{*}$-algebra-valued b-metric space.

1 Introduction

The Banach contraction principle [1], also known as the Banach fixed point theorem, is one of the main pillars of the theory of metric fixed points. According to this principle, if T is a contraction on a Banach space X, then T has a unique fixed point in X. Many researchers investigated the Banach fixed point theorem in many directions and presented generalizations, extensions, and applications of their findings. Among them, Bakhtin [2] introduced a prominent generalization of the idea of a metric space, which is later used by Czerwick [3, 4]. They introduced and used the concept of real-valued b-metric space to establish certain fixed point results. The idea clearly is an extension of the metric space as follows from the following definition.

Definition 1.1

([5])

Let X be a nonempty set, and $b \in\mathbb{R}$ be such that $b \geq1$. A b-metric on X is a real-valued mapping $d_{b}\colon X \times X \rightarrow\mathbb{R}$ that satisfies the following conditions for all $x,y,z \in X$:

• (1) $d_{b}(x,y)\ge0\mbox{ and }d_{b}(x,y)= 0 \Leftrightarrow x=y$.

• (2) $d_{b}(y,x)=d(x,y)$ (symmetry).

• (3) $d_{b}(y,z)\le b [d_{b}(y,x)+d_{b}(x,z)]$.

By a b-metric space with coefficient b we mean the pair $(X, d_{b})$.

For recent development on b-metric spaces, we refer to [5–10].

Recently, Ma et al. [11] presented their work on the extension of Banach contraction principle for $C^{*}$-algebra-valued metric spaces. Later, Batul and Kamran [12] introduced the notion of a $C^{*}$-valued contractive type mapping and established a fixed point result in this setting. Motivated by the ideas and results presented in [11, 12], in this paper, we will introduce a new notion of $C^{*}$-algebra-valued b-metric space and establish a fixed point result in such spaces.

We now recollect some basic definitions, notation, and results. The details on $C^{*}$-algebras are available in [13, 14].

An algebra $\mathbb{A}$, together with a conjugate linear involution map $a\mapsto a^{*}$, is called a ∗-algebra if $(ab)^{*}=b^{*}a^{*}$ and $(a^{*})^{*}=a$ for all $a,b \in\mathbb{A}$. Moreover, the pair $(\mathbb{A},*)$ is called a unital ∗-algebra if $\mathbb{A}$ contains the identity element $1_{\mathbb{A}}$. By a Banach ∗-algebra we mean a complete normed unital ∗-algebra $(\mathbb{A},*)$ such that the norm on $\mathbb{A}$ is submultiplicative and satisfies $\|a^{*} \|=\|a \|$ for all $a\in\mathbb{A}$. Further, if for all $a\in\mathbb{A}$, we have $\|a^{*}a \|=\|a \|^{2}$ in a Banach ∗-algebra $(\mathbb{A}, *)$, then $\mathbb{A}$ is known as a $C^{*}$-algebra. A positive element of $\mathbb{A}$ is an element $a \in\mathbb{A}$ such that $a=a^{*}$ and its spectrum $\sigma(a)\subset\mathbb{R_{+}}$, where $\sigma(a)=\lbrace\lambda \in\mathbb{R} : \lambda1_{\mathbb{A}}\mbox{-}a \mbox{ is noninvertible}\rbrace$. The set of all positive elements will be denoted by $\mathbb{A}_{+}$. Such elements allow us to define a partial ordering ‘⪰’ on the elements of $\mathbb{A}$. That is,

$$b \succeq a \quad\mbox{if and only if}\quad b-a \in\mathbb{A}_{+}. $$

If $a\in\mathbb{A}$ is positive, then we write $a \succeq 0_{\mathbb{A}}$, where $0_{\mathbb{A}}$ is the zero element of $\mathbb{A}$. Each positive element a of a $C^{*}$-algebra $\mathbb{A}$ has a unique positive square root. From now on, by $\mathbb{A}$ we mean a unital $C^{*}$-algebra with identity element $1_{\mathbb{A}}$. Further, $\mathbb{A}_{+} = \lbrace a\in\mathbb{A}:a\succeq0_{\mathbb{A}} \rbrace$ and $(a^{*}a)^{1/2}=\vert a \vert$. Using the concept of positive elements in $\mathbb{A}$, a $C^{*}$-algebra-valued metric d on a nonempty set X is defined in [11] as a mapping $d\colon X\times X \rightarrow\mathbb{A}_{+}$ that satisfies, for all $x_{1},x_{2},x_{3} \in X$, (i) $d(x_{1},x_{2})=0_{\mathbb{A}} \Leftrightarrow x_{1}=x_{2}$, (ii) $d(x_{1},x_{2})=d(x_{2},x_{1})$, and (iii) $d(x_{1},x_{2})\preceq d(x_{1},x_{3})+d(x_{3},x_{2})$. The triplet $(X,\mathbb{A},d)$ is then called a $C^{*}$-algebra-valued metric space.

2 Main results

In this section, we extend Definition 1.1 to introduce the notion b-metric space in the setting of $C^{*}$-algebras as follows.

Definition 2.1

Let $\mathbb{A}$ be a $C^{*}$-algebra, and X be a nonempty set. Let $b \in\mathbb{A}$ be such that $\|b \| \geq1$. A mapping $d_{b}\colon X \times X \rightarrow\mathbb{A}_{+}$ is said to be a $C^{*}$-algebra-valued b-metric on X if the following conditions hold for all $x_{1},x_{2},x_{3} \in\mathbb{A}$:

• (BM1) $d_{b}(x_{1},x_{2})=0_{\mathbb{A}} \Leftrightarrow x_{1}=x_{2}$.

• (BM2) $d_{b}$ is symmetric, that is, $d_{b}(x_{1},x_{2})=d_{b}(x_{2},x_{1})$.

• (BM3) $d_{b}(x_{1},x_{2})\preceq b [d_{b}(x_{1},x_{3})+d_{b}(x_{3},x_{2})]$.

The triplet $(X,\mathbb{A}, d_{b})$ is called a $C^{*}$-algebra-valued b-metric space with coefficient b.

Remark 2.1

Note that:

• (1) If we take $\mathbb{A}=\mathbb{R}$, then the new notion of $C^{*}$-algebra-valued b-metric space becomes equivalent to Definition 1.1 of the real b-metric space.

• (2) If we take $b=1_{\mathbb{A}}$ in Definition 2.1, then $d_{b}$ becomes the usual $C^{*}$-algebra-valued metric as defined in [11].

Thus, the class of ordinary $C^{*}$-algebra-valued metric spaces is clearly smaller than the class of $C^{*}$-algebra-valued b-metric spaces. In fact, there are $C^{*}$-algebra-valued b-metric spaces that are not $C^{*}$-algebra-valued metric spaces, as illustrated by the following example.

Example 2.1

Let $X=\ell_{p}$ be the set of sequences $\{x_{n}\}$ in $\mathbb{R}$ such that $\sum_{n=1}^{\infty}|x_{n}|^{p} < \infty$ and $0< p<1$. Let $\mathbb {A}=M_{2}(\mathbb{R})$. For $x=x_{n}, y=y_{n} \in\ell_{p}$, define $d_{b}:X \times X \rightarrow \mathbb{A}$ as follows:

$$d_{b}(x,y) = \begin{pmatrix} (\sum_{n=1}^{\infty}|x_{n}-y_{n}|^{p} )^{\frac{1}{p}} & 0 \\ 0 & (\sum_{n=1}^{\infty}|x_{n}-y_{n}|^{p} )^{\frac{1}{p}} \end{pmatrix}. $$

Then one can show that $d_{b}$ is a $C^{*}$-algebra-valued b-metric space with coefficient $b =\bigl( {\scriptsize\begin{matrix}{} 2^{\frac{1}{p}} & 0 \cr 0 & 2^{\frac{1}{p}} \end{matrix}}\bigr)$ such that $\|b\|=2^{\frac{1}{p}}$. The claim follows from the following observation in [4]:

$$ \Biggl(\sum_{n=1}^{\infty}|x_{n}-z_{n}|^{p} \Biggr)^{\frac{1}{p}} \le 2^{\frac{1}{p}} \Biggl[ \Biggl(\sum _{n=1}^{\infty}|x_{n}-y_{n}|^{p} \Biggr)^{\frac {1}{p}} + \Biggl(\sum_{n=1}^{\infty}|y_{n}-z_{n}|^{p} \Biggr)^{\frac{1}{p}} \Biggr]. $$

Note that here $d_{b}$ is not a usual $C^{*}$-algebra-valued metric on X.

From now on, we call a $C^{*}$-algebra-valued b-metric space simply a $C^{*}$-valued b-metric, and the triplet $(X,\mathbb{A},d_{b})$ is then called a $C^{*}$-valued b-metric space. Given $(X,\mathbb{A},d_{b})$, the following are natural deductions from the corresponding notions in $C^{*}$-valued metric spaces.

• (1) A sequence $\lbrace x_{n} \rbrace$ in X is said to be convergent to a point $x \in X$ with respect to the algebra $\mathbb{A}$ if and only if for any $\epsilon>0$, there is an $N \in\mathbb{N}$ such that $\|d_{b}(x_{n},x) \| < \epsilon$ for all $n> N$. Symbolically, we then write $\lim_{n\rightarrow \infty} x_{n}=x$.

• (2) If for any $\epsilon>0$, there exists $N \in\mathbb{N}$ such that $\|d_{b}(x_{n},x_{m}) \| < \epsilon$ for all $n, m > N$, then the sequence $\lbrace x_{n} \rbrace$ is called a Cauchy sequence with respect to $\mathbb{A}$.

• (3) If every Cauchy sequence in X is convergent with respect to $\mathbb{A}$, then the triplet $(X,\mathbb{A},d)$ is called a complete $C^{*}$-valued b-metric space.

Definition 2.2

Let $(X,\mathbb{A}, d_{b})$ be a $C^{*}$-valued b-metric space. A contraction on X is a mapping $T\colon X \rightarrow X$ if there exists an $a\in\mathbb{A}$ with $\| a \| < 1$ such that

$$ d_{b}(Tx,Ty)\preceq a^{*}d_{b}(x,y)a \quad\mbox{for all } x,y \in X. $$

(1)

Example 2.2

Let $\mathbb{A}= \mathbb{R}^{2}$ and $X=[0,\infty)$. Let ⪯ be the partial order on $\mathbb{A}$ given by

$$\begin{aligned}& (a_{1},b_{1})\preceq(a_{2},b_{2}) \quad\Leftrightarrow\quad a_{1} \leq a_{2} \mbox{ and } b_{1} \leq b_{2}. \end{aligned}$$

Define

$$d_{b}\colon X \times X \rightarrow\mathbb{A},\qquad d_{b}(x,y)= \bigl((x-y)^{2},0\bigr). $$

Then $d_{b}$ is $C^{*}$-valued b-metric with coefficient $(2,0)$, and with this $d_{b}$, the triplet $(X,\mathbb{A},d_{b})$ becomes a $C^{*}$-valued b-metric. Consider $T\colon X \rightarrow X$ given by $Tx=\frac{x}{3}+5$; then T is a contraction on X with $a=(\frac{1}{3},0)$:

$$\begin{aligned}[b] d_{b}(Tx,Ty)= \bigl((Tx-Ty)^{2},0 \bigr) = \biggl( \biggl(\frac{x}{3}-\frac{y}{3} \biggr)^{2},0 \biggr) = \biggl(\frac{1}{3},0 \biggr)d_{b}(x,y) \biggl( \frac{1}{3},0 \biggr). \end{aligned} $$

Theorem 2.1

Consider a complete $C^{*}$-valued b-metric space $(X,\mathbb{A},d_{b})$ with coefficient b. Let $T\colon X \rightarrow X$ be a contraction with the contraction constant a such that $\| b\| \|a \|^{2} < 1$. Then T has a unique fixed point in X.

Proof

If $\mathbb{A} = \{0_{\mathbb{A}}\}$, then there is nothing to prove. Assume that $\mathbb{A}\ne\{0_{\mathbb{A}}\}$.

Choose $x_{0} \in X$ and define inductively a sequence $\{x_{n}\}$ by the iterative scheme as

$$x_{n+1}=Tx_{n}. $$

Then it follows that $x_{n}=T^{n}x_{0}$ for $n=0,1,2, \ldots$ . From the contraction condition (1) on T it follows that

$$\begin{aligned} d_{b}(x_{n},x_{n+1}) =& d_{b}(Tx_{n-1},Tx_{n}) \\ \preceq& a^{*}d_{b}(x_{n-1},x_{n})a \\ =& a^{*}d_{b}(Tx_{n-2},Tx_{n-1})a \\ \preceq& \bigl(a^{*}\bigr)^{2}d_{b}(x_{n-2},x_{n-1})a^{2} \\ \preceq& \bigl(a^{*}\bigr)^{3}d_{b}(x_{n-3},x_{n-2})a^{3} \preceq \bigl(a^{*}\bigr)^{n}d_{b}(x_{0},x_{1})a^{n}= \bigl(a^{*}\bigr)^{n}Da^{n}, \end{aligned}$$

where $D=d_{b}(x_{0},x_{1})$.

Now suppose that $m>n$; then the triangle inequality (BM3) for the b-metric $d_{b}$ implies

$$\begin{aligned} d_{b}(x_{n},x_{m}) \preceq& b d(x_{n},x_{n+1}) + b^{2}d(x_{n+1},x_{n+2})+ \cdots+ b^{m-n-1}d(x_{m-2},x_{m-1}) \\ &{}+ b^{m-n-1}d(x_{m-1},x_{m}) \\ \preceq& b\bigl(a^{*}\bigr)^{n}Da^{n} +b^{2} \bigl(a^{*}\bigr)^{n+1}Da^{n+1} + \cdots+ b^{m-n-1} \bigl(a^{*}\bigr)^{m-2}Da^{m-2} \\ &{}+ s^{m-n-1}\bigl(a^{*}\bigr)^{m-1}Da^{m-1} \\ =& b\bigl[\bigl(a^{*}\bigr)^{n}Da^{n} +b\bigl(a^{*} \bigr)^{n+1}Da^{n+1} + \cdots+ b^{m-n-2}\bigl(a^{*} \bigr)^{m-2}Da^{m-2}\bigr] \\ &{}+ b^{m-n-1}\bigl(a^{*}\bigr)^{m-1}Da^{m-1} \\ =& b\sum_{k=n}^{m-2}b^{k-n} \bigl(a^{*}\bigr)^{k}Da^{k} + b^{m-n-1}\bigl(a^{*} \bigr)^{m-1}Da^{m-1} \\ =& b\sum_{k=n}^{m-1}b^{k-n} \bigl(a^{*}\bigr)^{k}D^{\frac{1}{2}}D^{\frac{1}{2}}a^{k} + b^{m-n-1}\bigl(a^{*}\bigr)^{m-1}D^{\frac{1}{2}}D^{\frac{1}{2}}a^{m-1} \\ =& b\sum_{k=n}^{m-1}b^{k-n} \bigl(D^{\frac{1}{2}}a^{k}\bigr)^{*} \bigl(D^{\frac{1}{2}}a^{k} \bigr) + b^{m-n-1}\bigl(D^{\frac{1}{2}}a^{m-1}\bigr)^{*} \bigl(D^{\frac{1}{2}}a^{m-1}\bigr) \\ =& b\sum_{k=n}^{m-1}b^{k-n}\bigl|D^{\frac{1}{2}}a^{k}\bigr|^{2} + b^{m-n-1}\bigl|D^{\frac {1}{2}}a^{m-1}\bigr|^{2} \\ \preceq& \Biggl\| b\sum_{k=n}^{m-1}b^{k-n}\bigl|D^{\frac{1}{2}}a^{k}\bigr|^{2} \Biggr\| 1_{\mathbb{A}} + \bigl\| b^{m-n-1}\bigl|D^{\frac{1}{2}}a^{m-1}\bigr|^{2} \bigr\| 1_{\mathbb{A}} \\ \preceq& \|b\|\sum_{k=n}^{m-1} \bigl\| b^{k-n}\bigr\| \bigl\| D^{\frac{1}{2}}\bigr\| ^{2} \bigl\| a^{k} \bigr\| ^{2} 1_{\mathbb{A}} + \bigl\| b^{m-n-1}\bigr\| \bigl\| D^{\frac{1}{2}} \bigr\| ^{2} \bigl\| a^{m-1}\bigr\| ^{2} 1_{\mathbb{A}} \\ \preceq& \|b\|\sum_{k=n}^{m-1} \|b\|^{k-n} \bigl\| D^{\frac{1}{2}}\bigr\| ^{2} \bigl\| a^{k} \bigr\| ^{2} 1_{\mathbb{A}} + \|b\|^{m-n-1} \bigl\| D^{\frac{1}{2}} \bigr\| ^{2} \bigl\| a^{m-1}\bigr\| ^{2} 1_{\mathbb{A}} \\ \preceq& \|b\|^{1-n} \bigl\| D^{\frac{1}{2}}\bigr\| ^{2}\sum _{k=n}^{m-1}\|b\|^{k} \bigl\| a^{2}\bigr\| ^{k} 1_{\mathbb{A}} + \|b\|^{-n}\|b \|^{m-1} \bigl\| D^{\frac{1}{2}}\bigr\| ^{2} \bigl\| a^{m-1} \bigr\| ^{2} 1_{\mathbb{A}} \\ \preceq& \|b\|^{1-n} \bigl\| D^{\frac{1}{2}}\bigr\| ^{2}\sum _{k=n}^{m-1}\bigl(\|b\| \bigl\| a^{2}\bigr\| \bigr)^{k} 1_{\mathbb{A}} + \|b\|^{-n}\bigl\| D^{\frac{1}{2}} \bigr\| ^{2}\bigl(\|b\| \bigl\| a^{2}\bigr\| \bigr)^{m-1} 1_{\mathbb{A}} \\ \longrightarrow& 0_{\mathbb{A}} \quad\mbox{as } m, n \rightarrow\infty, \end{aligned}$$

which follows from the observation that the summation in the first term is a geometric series, and $\|b\|\|a^{2}\| < 1$ implies that both $(\|b\| \|a^{2}\|)^{m-1} \rightarrow0$ and $(\|b\| \|a^{2}\|)^{n-1} \rightarrow0$. This proves that $\{x_{n}\}$ is a Cauchy sequence in X with respect to $\mathbb{A,}$ and from the completeness of $(X, \mathbb{A}, d)$ it follows that $x_{n} \rightarrow x \in X$, that is,

$$\lim_{n\rightarrow\infty} x_{n} = \lim_{n\rightarrow\infty} Tx_{n-1} = x . $$

We claim that x is a fixed point of T. In fact, from the triangle inequality (BM3) and the contraction condition (1) we have:

$$\begin{aligned} 0_{\mathbb{A}} \preceq& d(Tx,x) \\ \preceq& b\bigl[d(Tx,Tx_{n})+d(Tx_{n},x)\bigr] \\ \preceq& b a^{*}d(x,x_{n})a + d(x_{n-1},x) \longrightarrow 0_{\mathbb{A}} \quad\mbox{as } n\rightarrow\infty. \end{aligned}$$

This shows that $Tx=x$.

To prove that x is the unique fixed point, we suppose that $y\in X$ is another fixed point of T. Then again from the contraction condition (1) we have

$$\begin{aligned} 0_{\mathbb{A}} \preceq d(x,y) = d(Tx,Ty) \preceq a^{*} d(x,y) a. \end{aligned}$$

Using the norm of $\mathbb{A}$, we have

$$\begin{aligned} 0\le\bigl\| d(x,y)\bigr\| \le\bigl\| a^{*} d(x,y) a\bigr\| \le\bigl\| a^{*}\bigr\| \bigl\| d(x,y)\bigr\| \|a\| =\|a\| ^{2} \bigl\| d(x,y)\bigr\| . \end{aligned}$$

The above inequality holds only when $d(x,y) = 0_{\mathbb{A}}$. Hence, $x=y$. □

Example 2.3

The mapping T of Example 2.2 satisfies the hypothesis of Theorem 2.1, and T has unique fixed point $x=1.5$ in X.

Remark 2.2

Theorem 2.1 generalizes the following results.

• (1) By taking $\mathbb{A} =\mathbb{R}$, the $C^{*}$-valued b-metric becomes simply the b-metric, and we immediately get the Banach contraction principle in b-metric spaces from Theorem 2.1.

• (2) Taking $b=1$, [11], Theorem 2.1, becomes a special case of Theorem 2.1.

3 Application

As an application of the fixed point theorem for contractions on a $C^{*}$-valued complete b-metric space, we provide an existence result for a class of integral equations.

Example 3.1

Let E be a Lebesgue-measurable set and $X=L^{\infty}(E)$. Consider the Hilbert space $L^{2}(E)$. Let the set of all bounded linear operators on $L^{2}(E)$ be denoted by $BL(L^{2}(E))$. Note that $BL(L^{2}(E))$ is a $C^{*}$-algebra with usual operator norm. For $S, T \in X$, define

$$d_{b}\colon X \times X \rightarrow BL\bigl(L^{2}(E)\bigr),\qquad d_{b}(T,S)=\pi_{(T-S)^{2}}, $$

where $\pi_{h}\colon L^{2}(E)\rightarrow L^{2}(E)$ is the product operator given by

$$\pi_{h}(f)=h\cdot f \quad\mbox{for } f \in L^{2}(E). $$

Working in the same lines as in [11], Example 2.1, we can show that $(X,BL(L^{2}(E)),d_{b})$ is a complete $C^{*}$-valued b-metric space. With these settings, suppose that there exist a continuous function $f \colon E\times E \rightarrow\mathbb{R}$ and a constant $0< \alpha<1$ such that for all $x, y \in X$ and $u,v \in E$, we have

$$ \bigl|K\bigl(u,v, x(v)\bigr) - K\bigl(u, v, y(v)\bigr)\bigr| \le\alpha\bigl|f(u,v) \bigl(x(v)-y(v)\bigr)\bigr|, $$

(2)

where K is a function from $E \times E \times\mathbb{R}$ to $\mathbb{R}$, and $\sup_{t\in E} \int_{E} |f (u,v)|\,dv \le1$. Then the integral equation

$$x(u)= \int_{E} K\bigl(u,v,x(v)\bigr)\,dv,\quad u\in E $$

has a unique solution.

Proof

Here $(X,BL(L^{2}(E)),d_{b})$ is a $C^{*}$-valued complete b-metric space with respect to $BL(L^{2}(E))$.

Let

$$T\colon X\rightarrow X,\quad Tx(u)= \int_{E} K\bigl(u,v,x(v)\bigr)\,dv,\quad u\in E. $$

Then

$$\begin{aligned} \bigl\| d(Tx,Ty)\bigr\| =& \|\pi_{(Tx-Ty)^{2}}\| \\ =& \sup_{\|g\|=1} \langle\pi_{(Tx-Ty)^{2}}g,g\rangle\quad \mbox{for every } g\in L^{2}(E) \\ =& \sup_{\|g\|=1} \int_{E} (Tx-Ty)^{2}g(u)\overline{g(u)}\,dv \\ =& \sup_{\|g\|=1} \int_{E} \biggl[ \int_{E} \bigl(K\bigl(u,v,x(v)\bigr)-K\bigl(u,v,y(v)\bigr) \bigr)\,dv \biggr]^{2} g(u)\overline{g(u)}\,du \\ \le& \sup_{\|g\|=1} \int_{E} \biggl[ \int_{E} \bigl(K\bigl(u,v,x(v)\bigr)-K\bigl(u,v,y(v)\bigr) \bigr)\,dv \biggr]^{2}\bigl|g(u)\bigr|^{2}\,du \\ \le& \sup_{\|g\|=1} \int_{E} \alpha^{2} \biggl[ \int_{E} \bigl(f(u,v) \bigl(x(v)-y(v)\bigr)\bigr)\,dv \biggr]^{2}\bigl|g(u)\bigr|^{2}\,du \\ \le& \alpha^{2} \sup_{\|g\|=1} \int_{E} \biggl[ \int_{E} \bigl|f(u,v)\bigr|\,dv \biggr]^{2}\bigl|g(u)\bigr|^{2}\,du \cdot\bigl\| (x-y)^{2}\bigr\| _{\infty}\\ \le& \alpha^{2} \sup_{t \in E} \int_{E} \bigl|f(u,v)\bigr|^{2}\,dv \cdot\sup _{\|g\|=1} \int_{E} \bigl|g(u)\bigr|^{2}\,du \cdot\bigl\| (x-y)^{2} \bigr\| _{\infty}\\ \le& \alpha^{2}\bigl\| (x-y)^{2}\bigr\| _{\infty}\\ = & \|a\| \bigl\| d(x,y)\bigr\| . \end{aligned}$$

Setting $a= \alpha I_{2}$, we have $a\in BL(L^{2}(E))_{+}$ and $\|a\|=\alpha^{2} <1$. Thus, all the conditions of Theorem 2.1 hold, and hence the conclusion. □