## Abstract

The paper is focused on general sequences of discrete linear operators, say \((L_{n})_{n}\geq1\). The special case of positive operators is also to our attention. Concerning the quantity \(\triangle(L_{n},f,g):=L_{n}(f_{g})-(L_{n}f)(L_{n}g)\),\(\ f\) and g belonging to some certain spaces, we propose different estimates. Firstly, we study its asymptotic behavior in Voronovskaja’s sense. Examples are presented. Secondly, we prove an extension of Chebyshev-Gruss type inequality for the above quantity. Special cases are investigated separately. Finally we establish sufficient conditions that ensure statistical convergence of the sequence \(\triangle(L_{n},f,g)\).

## Authors

**Octavian Agratini**

Department of Mathematics, Babes-Bolyai University, Cluj-Napoca, Romania

## Keywords

Linear operator ·Voronovskaja formula · Grüss-type inequality · Bernstein operator · Jain operator · Generalized sample operator · Statistical convergence

## Paper coordinates

O. Agratini, *Properties of discrete non-multiplicative operators*, Analysis and Mathematical Physics, 9 (2020), pp. 131-141. https://doi.org/10.1007/s13324-017-0186-4

requires subscription: https://doi.org/10.1007/s13324-017-0186-4

## About this paper

##### Print ISSN

?

##### Online ISSN

google scholar link

## Paper (preprint) in HTML form

# Properties of discrete non-multiplicative operators

###### Abstract.

The paper is focused on general sequences of discrete linear operators, say ${({L}_{n})}_{n\ge 1}$. The special case of positive operators is also to our attention. Concerning the quantity $\mathrm{\Delta}({L}_{n},f,g):={L}_{n}(fg)-({L}_{n}f)({L}_{n}g)$, $f$ and $g$ belonging to some certain spaces, we propose different estimates. Firstly, we study its asymptotic behavior in Voronovskaja’s sense. Examples are presented. Secondly, we prove an extension of Chebyshev-Grüss type inequality for the above quantity. Special cases are investigated separately. Finally we establish sufficient conditions that ensure statistical convergence of the sequence $\mathrm{\Delta}({L}_{n},f,g)$.

Keywords and phrases: Linear operator, Voronovskaja formula, Grüss-type inequality, Bernstein operator, Jain operator, generalized sample operator, statistical convergence.

Mathematics Subject Classification: 47A63, 41A25.

## 1. Introduction

A current subject in Approximation Theory is the approximation of functions by positive linear operators. Usually, two types of positive approximation processes are used – the discrete respectively continuous form. In the first case, various classes of operators are expressed by series. In this paper we are interested in these last sequences, denoted generically by ${({L}_{n})}_{n\ge 1}$. The structure of these operators is detailed in Section 2. Since a property often seen in linear positive operators is to reproduce the constants which are involved in preservation of the monomial ${e}_{0}$, ${e}_{0}(x)=1$, we assume that the following relation

$${L}_{n}{e}_{0}={e}_{0},n\in \mathbb{N},$$ | (1.1) |

occurs.

In recent years, it was studied the behavior of the expression

$${L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x)$$ | (1.2) |

for given functions $f$, $g$ belonging to the domain of ${L}_{n}$ and $x$ fixed and compatible with the domain of the functions. The analysis of quantity (1.2) took different approaches, but usually on particular cases of linear (positive or non-positive) operators.

In the present paper we deal with the following two aspects.

Among the properties studied for linear approximation processes, the asymptotic behavior known as the so called Voronovskaja type formula plays an important role. Knowing a such Voronovskaja formula verified by ${L}_{n}$, in Section 3 we establish the asymptotic behavior of quantity defined by (1.2). Particular cases are shown.

In Section 4 we present a Chebyshev-Grüss-type inequality involving linear operators without supposing that they are positive. Also, we indicate sufficient conditions that ensure the statistical convergence of the expression given by (1.2).

## 2. Construction of the operators ${L}_{n}$

Let $J\subset \mathbb{R}$ be an interval and its interior $Int(J)=(a,b)$, $$. Since a linear substitution maps a bounded interval $(a,b)$ into the interval $(0,1)$, without loss of generality, we will consider as $Int(J)$ one of the intervals $(0,1)$, $(0,\mathrm{\infty})={\mathbb{R}}_{+}^{\ast}$, $(-\mathrm{\infty},\mathrm{\infty})=\mathbb{R}$.

Set ${\mathbb{N}}_{0}=\{0\}\cup \mathbb{N}$. Let ${\mathrm{\Delta}}_{n}={({x}_{n,k})}_{k\in {I}_{n}}$ be a net on the interval $J$, where ${I}_{n}\subseteq \mathbb{Z}$ is a set of indices consistent with $J$, this meaning $\{{x}_{n,k}:k\in {I}_{n}\}\subset J$. We denote by $B(J)$, respectively $C(J)$, the space of all real-valued bounded, respectively continuous, functions defined by $J$.

Linear operators of discrete type are often designed as follows

$$({L}_{n}f)(x)=\sum _{k\in {I}_{n}}{a}_{n,k}(x)f({x}_{n,k}),n\in \mathbb{N},x\in J,$$ | (2.1) |

where ${a}_{n,k}\in C(J)$, $k\in {I}_{n}$, such that

$$ | (2.2) |

and

$$\sum _{k\in {I}_{n}}{a}_{n,k}(x)=1.$$ | (2.3) |

The last condition ensures the fulfillment of relation (1.1). In order that ${L}_{n}$ to be also positive, we impose

$${a}_{n,k}\ge 0,k\in {I}_{n}.$$ | (2.4) |

In this case, ${({a}_{n,k})}_{k\in {I}_{n}}$ forms a so called blending system. The operator ${L}_{n}$ being linear and positive, is continuous with its norm equal to 1, because of (1.1) takes place.

The domain of the operators ${L}_{n}$, $n\in \mathbb{N}$, denoted by $\mathcal{F}(J)$, consists of the set of all functions for which the right-hand side of the relation (2.1) is well defined. For example, due to (2.3), ${e}_{0}\in \mathcal{F}(J)$. Moreover, based on (2.2), we get $B(J)\subset \mathcal{F}(J)$,

Regarding the asymptotic behavior of linear (not necessarily positive) operators ${L}_{n}$, we make the following assumption. A real constant $\lambda >0$ and a second order differential operator $\varphi $ exist such that

$$\underset{n\to \mathrm{\infty}}{lim}{n}^{\lambda}(({L}_{n}f)(x)-f(x))=\varphi (f,x),$$ | (2.5) |

for some $x\in J$, where

$$\varphi (\cdot ,x)=\alpha (x)\frac{d}{dx}+\beta (x)\frac{{d}^{2}}{d{x}^{2}},$$ |

$\alpha $, $\beta $ being continuous functions defined on $J$.

In order to take place relation (2.5), function $f$ has to be differentiable in some neighborhood of $x$ and has second derivative ${f}^{\prime \prime}(x)$. This is a Voronovskaja type result given in its local version.

Here we give some examples of formula described by (2.5) associated with well known linear positive operators. The cases $J=[0,1]$, $J={\mathbb{R}}_{+}$, $J=\mathbb{R}$ are illustrated.

1. Bernstein operators. $J=[0,1]$, ${I}_{n}=\{0,1,\mathrm{\dots},n\}$, ${x}_{n,k}={\displaystyle \frac{k}{n}}$,

$$({L}_{n}f)(x)\equiv ({B}_{n}f)(x)=\sum _{k=0}^{n}\left(\genfrac{}{}{0pt}{}{n}{k}\right){x}^{k}{(1-x)}^{n-k}f\left(\frac{k}{n}\right).$$ | (2.6) |

Set ${C}^{2}([0,1])$ the space twice continuously differentiable functions defined on $[0,1]$. It takes place [17]

$$\underset{n\to \mathrm{\infty}}{lim}n(({B}_{n}f)(x)-f(x))=\frac{x(1-x)}{2}{f}^{\prime \prime}(x),f\in {C}^{2}([0,1]).$$ |

The above convergence is uniform on $[0,1]$. Here $\lambda =1$, $\alpha (x)=0$ and $\beta (x)=x(1-x)/2$.

We notice that Voronovskaja gave the first asymptotic formula for the pointwise approximation of continuous functions which have a second derivative at a point $x\in J=[0,1]$.

2. Jain operators. $J=[0,\mathrm{\infty})$, ${I}_{n}={\mathbb{N}}_{0}$, ${x}_{n,k}={\displaystyle \frac{k}{n}}$,

$$({L}_{n}f)(x)\equiv ({P}_{n}^{[\beta ]}f)(x)=\sum _{k=0}^{\mathrm{\infty}}{\omega}_{\beta}(k,nx)f\left(\frac{k}{n}\right),$$ | (2.7) |

where

$${\omega}_{\beta}(k,\alpha )=\frac{\alpha {(\alpha +k\beta )}^{k-1}}{k!}{e}^{-(\alpha +k\beta )},k\in {\mathbb{N}}_{0},$$ |

whenever the sum in (2.7) is convergent. Jain [11] showed that for $\alpha >0$ and $$,

$$\sum _{k=0}^{\mathrm{\infty}}{\omega}_{\beta}(k,\alpha )=1.$$ |

Note that this is a probability distribution which is called generalized Poisson distribution.

The special case $\beta =0$, Jain operators reduce to the classical Szász-Mirakjan operators [15], [12].

In (2.7) the parameter $\beta $ may depend on a natural number $n$, say $\beta :={\beta}_{n}$. If $$, $\underset{n\to \mathrm{\infty}}{lim}{\beta}_{n}=0$, and a suitable condition on the growth of $f$ is satisfied, then Farcaş [5] proved

$$\underset{n\to \mathrm{\infty}}{lim}n(({P}_{n}^{[{\beta}_{n}]}f)(x)-f(x))=\frac{x}{2}{f}^{\prime \prime}(x),x>0.$$ |

Here $\lambda =1$, $\alpha (x)=0$, $\beta (x)=x/2$.

We point out that the same Voronovskaja formula is verified by genuine Szász-Mirakjan operators.

3. Sampling operators. $J=\mathbb{R}$, ${I}_{n}=\mathbb{Z}$, ${x}_{n,k}={\displaystyle \frac{k}{n}}$.

We consider a compactly supported function $\phi \in C(\mathbb{R})$ satisfying the following assumptions

$$ |

$$\sum _{k\in \mathbb{Z}}(k-x)\phi (x-k)=0,\sum _{k\in \mathbb{Z}}{(k-x)}^{2}\phi (x-k)=C,$$ | (2.8) |

for a given constant $C\in \mathbb{R}$ and every $x\in \mathbb{R}$. Following the results obtained by Butzer, Stens, Ries, see e.g., [3], [13], the sampling operators are defined by

$$({G}_{n}f)(x)=\sum _{k\in \mathbb{Z}}\phi (nx-k)f\left(\frac{k}{n}\right).$$ |

Since $\phi $ has a compact support, for each fixed $n\in \mathbb{N}$ and $x\in \mathbb{R}$, the above sum has a finite number of terms. Consequently, we can associate ${G}_{n}$ operators to any real-valued function $f$ defined on the whole real line.

Set ${L}^{\mathrm{\infty}}(\mathbb{R})$, the space of all real-valued functions defined on $\mathbb{R}$, essentially bounded and measurable. For any function $f\in {L}^{\mathrm{\infty}}(\mathbb{R})$ and at every point $x\in \mathbb{R}$ in which ${f}^{\prime \prime}(x)$ exists, one has (see [2, Corollary 1])

$$\underset{n\to \mathrm{\infty}}{lim}{n}^{2}(({G}_{n}f)(x)-f(x))=\frac{C}{2}{f}^{\prime \prime}(x),$$ |

where $C$ is defined at (2.8). In this example $\lambda =2$, $\alpha (x)=0$ and $\beta (x)=C/2$.

It is worthy to note that in the paper of Bardaro and Mantellini [2] is given a Voronovskaja type formula of a general class of discrete operators. A distinct section represents a survey on various asymptotic formulas for many classical operators which can be deduced as special cases of the original outcome established in the quoted paper.

##
3. Voronovskaja’s formula for non-multiplicative

linear operators

We are only concerned with sequences ${({L}_{n})}_{n\ge 1}$ of linear operators forming an approximation process. This means that a subspace $\mathcal{A}(J)\subset \mathcal{F}(J)$ exists such that

$$\underset{n\to \mathrm{\infty}}{lim}({L}_{n}h)(x)=h(x),x\in J,h\in \mathcal{A}(J).$$ | (3.1) |

The above convergence is generally pointwise.

About Voronovskaja’s formula for quantity defined by (1.2), our result will be read as follows.

###### Theorem 3.1.

Let ${L}_{n}$, $n\in \mathbb{N}$, be operators defined by (2.1) such that conditions (2.5) and (3.1) are fulfilled. For $x\in J$ and any functions $f$, $g$ belonging to $\mathcal{A}(J)$ differentiable in some neighborhood of $x$ and having the second derivative ${f}^{\prime \prime}(x)$ respectively ${g}^{\prime \prime}(x)$, the following identity

$$\underset{n\to \mathrm{\infty}}{lim}{n}^{\lambda}({L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x))=2\beta (x){f}^{\prime}(x){g}^{\prime}(x)$$ | (3.2) |

takes place.

Proof. Let $x\in J$ and the functions $f$, $g$ belong to $\mathcal{A}(J)$. For the sake of simplicity we introduce the following notations

${A}_{n}:$ | $={n}^{\lambda}({L}_{n}(fg,x)-(fg)(x))-\varphi (fg,x)$ | ||

$={n}^{\lambda}({L}_{n}(fg,x)-(fg)(x))$ | |||

$-\alpha (x)({f}^{\prime}(x)g(x)+f(x){g}^{\prime}(x))$ | |||

$-\beta (x)({f}^{\prime \prime}(x)g(x)+2{f}^{\prime}(x){g}^{\prime}(x)+f(x){g}^{\prime \prime}(x)),$ | |||

${B}_{n}:$ | $={n}^{\lambda}({L}_{n}(f,x)-f(x))-\varphi (f,x)$ | ||

$={n}^{\lambda}({L}_{n}(f,x)-f(x))-\alpha (x){f}^{\prime}(x)-\beta (x){f}^{\prime \prime}(x),$ | |||

${C}_{n}:$ | $={n}^{\lambda}({L}_{n}(g,x)-g(x))-\varphi (g,x)$ | ||

$={n}^{\lambda}({L}_{n}(g,x)-g(x))-\alpha (x){g}^{\prime}(x)-\beta (x){g}^{\prime \prime}(x).$ |

After some calculations we obtain

${A}_{n}$ | $-g(x){B}_{n}-({L}_{n}f)(x){C}_{n}+(\alpha (x){g}^{\prime}(x)+\beta (x){g}^{\prime \prime}(x))(f(x)-{L}_{n}(f,x))$ | |||

$+2\beta (x){f}^{\prime}(x){g}^{\prime}(x)={n}^{\lambda}({L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x)).$ | (3.3) |

Based on the relations (2.5) and (3.1) we have

$$\underset{n\to \mathrm{\infty}}{lim}{A}_{n}=0,\underset{n\to \mathrm{\infty}}{lim}{B}_{n}=0,\underset{n\to \mathrm{\infty}}{lim}{C}_{n}=0$$ |

and

$$\underset{n\to \mathrm{\infty}}{lim}(f(x)-{L}_{n}(f,x))=0.$$ |

Returning at (3) and passing to the limit with respect to $n$, we arrive at the desired result (3.2). The proof is ended. $\mathrm{\square}$

Note the following: although Voronovskaja’s formula of ${L}_{n}$ operators depends on function $\alpha $, the limit in (3.2) does not contain $\alpha $.

Returning to the examples presented in the previous section, under the assumption that the requirements of Theorem 3.1 are fulfilled, the identity given by (3.2) takes the following particular forms.

${1}^{\circ}$ Bernstein operators

$$\underset{n\to \mathrm{\infty}}{lim}n({B}_{n}(fg,x)-{B}_{n}(f,x){B}_{n}(g,x))=x(1-x){f}^{\prime}(x){g}^{\prime}(x).$$ |

${2}^{\circ}$ Jain operators

$$\underset{n\to \mathrm{\infty}}{lim}n({P}_{n}^{[{\beta}_{n}]}(fg,x)-{P}_{n}^{[{\beta}_{n}]}(f,x){P}_{n}^{[{\beta}_{n}]}(g,x))=x{f}^{\prime}(x){g}^{\prime}(x).$$ |

In the special case ${\beta}_{n}=0$, these operators turn to so called Szász-Mirakjan operators and be noticed that the second member of the identity (3.2) remains unchanged.

${3}^{\circ}$ Sampling operators

$$\underset{n\to \mathrm{\infty}}{lim}{n}^{2}({G}_{n}(fg,x)-{G}_{n}(f,x){G}_{n}(g,x))=C{f}^{\prime}(x){g}^{\prime}(x),$$ |

where $C$ is defined by (2.8).

## 4. A Chebyshev-Grüss-type inequality

The roots of inequality on which we investigate goes back to the classical works of Chebyshev [4] and Grüss [10].

The first author proved

$$\left|\frac{1}{b-a}{\int}_{a}^{b}f(x)g(x)\mathit{d}x-\frac{1}{{(b-a)}^{2}}{\int}_{a}^{b}f(x)\mathit{d}x{\int}_{a}^{b}g(x)\mathit{d}x\right|\le \frac{{(b-a)}^{2}}{12}\Vert {f}^{\prime}\Vert \Vert {g}^{\prime}\Vert ,$$ | (4.1) |

$f,g\in {C}^{1}([a,b])$ and $\parallel \cdot \parallel $ is the sup-norm on the Banach space $C([a,b])$.

The second author extended (4.1) to Riemann integrable functions defined on $[a,b]$, say $f$ and $g$, obtaining

$$\left|\frac{1}{b-a}{\int}_{a}^{b}f(x)g(x)\mathit{d}x-\frac{1}{{(b-a)}^{2}}{\int}_{a}^{b}f(x)\mathit{d}x{\int}_{a}^{b}g(x)\mathit{d}x\right|$$ |

$$\le \frac{1}{4}({M}_{f}-{m}_{f})({M}_{g}-{m}_{g}),$$ |

where ${M}_{h}=\underset{t\in [a,b]}{sup}|h(t)|$, ${m}_{h}=\underset{t\in [a,b]}{inf}|h(t)|$ for $h\in \{f,g\}$.

Recently, in many papers, modeling these inequalities for non-multiplicative linear operators have been highlighted. We mention as pioneering reference papers [1] and [14]. In these articles the authors have explored in depth linear (positive) operators acting on continuous functions defined on a compact metric space. Among subsequent papers we also recall [7], [8].

###### Theorem 4.1.

Let ${L}_{n}$, $n\in \mathbb{N}$, be linear operators defined by (2.1). For $x\in J$ and any functions $f$, $g$ belonging to $\mathcal{F}(J)$,

$${L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x)$$ |

$$ | (4.2) |

takes place.

Proof. Let $x\in J$ be arbitrarily fixed. Based on identity (2.3), for each $n\in \mathbb{N}$ we can write

${L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x)$ | ||||

$={\displaystyle \sum _{k\in {I}_{n}}}{a}_{n,k}(x)(fg)({x}_{n,k})-\left({\displaystyle \sum _{k\in {I}_{n}}}{a}_{n,k}(x)f({x}_{n,k})\right)\left({\displaystyle \sum _{j\in {I}_{n}}}{a}_{n,j}(x)g({x}_{n,j})\right)$ | ||||

$={\displaystyle \sum _{k\in {I}_{n}}}{a}_{n,k}(x)\left({\displaystyle \sum _{j\in {I}_{n}}}{a}_{n,j}(x)\right)f({x}_{n,k})g({x}_{n,k})$ | ||||

$-{\displaystyle \sum _{k\in {I}_{n}}}{\displaystyle \sum _{j\in {I}_{n}}}{a}_{n,k}(x){a}_{n,j}(x)f({x}_{n,k})g({x}_{n,j})$ | ||||

$={\displaystyle \sum _{\begin{array}{c}(k,j)\in {I}_{n}^{2}\\ k\ne j\end{array}}}{a}_{n,k}(x){a}_{n,j}(x)f({x}_{n,k})g({x}_{n,k})-{\displaystyle \sum _{\begin{array}{c}(k,j)\in {I}_{n}^{2}\\ k\ne j\end{array}}}{a}_{n,k}(x){a}_{n,j}(x)f({x}_{n,k})g({x}_{n,j})$ | ||||

$={\displaystyle \sum _{\begin{array}{c}(k,j)\in {I}_{n}^{2}\\ k\ne j\end{array}}}{a}_{n,k}(x){a}_{n,j}(x)f({x}_{n,k})(g({x}_{n,k})-g({x}_{n,j}))$ | ||||

$$ | (4.3) |

In the second sum we make the change of variable $j\leftrightarrow k$ and this becomes

$$\sum _{\begin{array}{c}(k,j)\in {I}_{n}^{2}\\ k>j\end{array}}{a}_{n,j}(x){a}_{n,k}(x)f({x}_{n,j})(g({x}_{n,j})-g({x}_{n,k})).$$ |

Returning at (4) we continue to write

$$ |

Reaching (4.2), the proof is completed. $\mathrm{\square}$

###### Remark 4.2.

If (2.4) holds, in other words the linear operators ${L}_{n}$, $n\in \mathbb{N}$, are also positive, and ${e}_{i}\in \mathcal{F}(J)$ (${e}_{i}(x)={x}^{i}$, $i=1,2$), clearly the right-hand side of (4.2) is positive, consequently we rediscover the classical inequality ${L}_{n}{e}_{2}\ge {L}_{n}^{2}{e}_{1}$ satisfied by positive linear operators enjoying the property (1.1). For the special case $\mathcal{F}(J)=C([0,1])$ this inequality is verified even if (1.1) (which coincides with to (2.3)) is replaced by weaker condition ${L}_{n}{e}_{0}\le {e}_{0}$, see [16, Lemma 1].

Under the assumption of Theorem 4.1, identity (4.2) implies the following relation

$$|{L}_{n}(fg,x)-{L}_{n}(f,x){L}_{n}(g,x)|$$ |

$$ | (4.4) |

###### Remark 4.3.

We ask if the inequality (4.4) is sharp. Since we can find a linear operator ${L}_{n}$ and some functions $f$, $g$, such that the equality holds, the answer is affirmative. For example, we choose

$$J=[0,1],n=1,f=g={e}_{1},{I}_{1}=\{0,1\},{x}_{1,k}=k\text{where}k\in \{0,1\},$$ |

and ${L}_{1}\equiv {B}_{1}$, Bernstein operators of first degree, see (2.6). Since the following identities are well known

$${B}_{n}{e}_{1}={e}_{1},{B}_{n}{e}_{2}={e}_{2}+\frac{{e}_{1}({e}_{0}-{e}_{1})}{n},n\in \mathbb{N},$$ |

we get $|{B}_{1}({e}_{2},x)-{B}_{1}^{2}({e}_{1},x)|=x(1-x)$, and

$${a}_{2,2}(x)={x}^{2},{a}_{2,1}(x)=2x(1-x),{a}_{2,0}(x)={(1-x)}^{2}.$$ |

After a few calculations we obtain that the two members of the inequality (4.4) are equal.

In what follows we bring into light some particular cases of relation (4.4).

Denoted by ${\mathrm{Lip}}_{M}\alpha $ the set of all Lipschitzian functions $f\in {\mathbb{R}}^{J}$ of order $\alpha $ $(0\le \alpha \le 1)$ having the same Lipschitz constant $M$ $(M>0)$, i.e.,

$$|f(u)-f(v)|\le M{|u-v|}^{\alpha},(u,v)\in J\times J,$$ |

we can enunciate

###### Corollary 4.4.

Let ${L}_{n}$, $n\in \mathbb{N}$, be linear operators defined by (2.1). For $x\in J$ and any functions $f$, $g$ belonging to $\mathcal{F}(J)\cap {\mathrm{Lip}}_{M}\alpha $,

$$ |

takes place.

As we stated in Section 2, $B(J)\subset \mathcal{F}(J)$. Also, for any $h\in B(J)$ the real constants $m(h)$, $M(h)$ exist such that $m(h)\le h(t)\le M(h)$, $t\in J$. Consequently

$$|h(u)-h(v)|\le M(h)-m(h):=\mathrm{\Delta}(h),$$ | (4.5) |

and taking in view (4.4), we get

###### Corollary 4.5.

It is worth mentioning, by using the notion of oscillation associated to a certain functional, a similar result was obtained in [9, Theorem 8].

We return to quantity (1.2). If ${L}_{n}$, $n\in \mathbb{N}$, are linear positive operators on the space $C(J)$, $J=[0,1]$, and (1.1) holds, then there is a Borel probability measure $\mu $ on $[0,1]$ such that ${L}_{n}h={\displaystyle {\int}_{0}^{1}}h\mathit{d}\mu $, for any $h\in C([0,1])$. We can write

$${L}_{n}{h}^{2}-{({L}_{n}h)}^{2}={\int}_{0}^{1}{h}^{2}\mathit{d}\mu -{\left({\int}_{0}^{1}h\mathit{d}\mu \right)}^{2}\ge 0.$$ |

This quantity represents the variance of the function $h$ with respect to Borel measure $\mu $. Set

$${V}_{n}(h,x)={L}_{n}({h}^{2},x)-{({L}_{n}(h,x))}^{2},x\in [0,1].$$ |

Our goal is to analyze the expression (1.2) in terms of the statistical convergence.

We pause to collect some notation and definitions.

The density of a set $U\subset \mathbb{N}$ is defined by

$$\delta (U)=\underset{n\to \mathrm{\infty}}{lim}\frac{1}{n}\sum _{k=1}^{n}{\chi}_{U}(k),$$ |

provided the limit exists, where ${\chi}_{U}$ is the characteristic function of $U$. A sequence of real numbers $x={({x}_{k})}_{k\ge 1}$ is statistically convergent to a real number $l$, denoted $st-\underset{k}{lim}{x}_{k}=l$, if, for every $\epsilon >0$,

$$\delta (\{k\in \mathbb{N}:|{x}_{k}-l|\ge \epsilon \})=0$$ |

holds. In Approximation Theory by linear positive operators, the statistical convergence has been examined for the first time in 2002 by A.D. Gadjiev and C. Orhan [6] who gave the Bohman-Korovkin criterion via statistical convergence.

###### Theorem 4.6.

Let ${L}_{n}:C([0,1])\to C([0,1])$, $n\in \mathbb{N}$, linear positive operators defined at (2.1) such that ${L}_{n}{e}_{0}\le {e}_{0}$. Let $f$, $g$ belong to $C([0,1])$ arbitrarily fixed such that one of these functions has its variance statistically convergent to zero. The following relation

$$st-\underset{n\to \mathrm{\infty}}{lim}\Vert {L}_{n}(fg)-({L}_{n}f)({L}_{n}g)\Vert =0$$ | (4.6) |

holds.

Proof. Let $f$ and $g$ in $C([0,1])$ be given. Consistent with the hypothesis, we can consider

$$st-\underset{n\to \mathrm{\infty}}{lim}\Vert {V}_{n}g\Vert =0.$$ | (4.7) |

Also, it is clear that $\Vert {V}_{n}f\Vert \le 2{\Vert f\Vert}^{2}$, the second term being independent of $n$ and $x\in [0,1]$.

Throughout this theorem, $\parallel \cdot \parallel $ represents the uniform norm of the space $C([0,1])$. The key relation used in the proof was established by Uchiyama [16, Eq. (3)], namely

$$\Vert {L}_{n}(fg)-({L}_{n}f)({L}_{n}g)\Vert \le {\Vert {V}_{n}f\Vert}^{1/2}{\Vert {V}_{n}g\Vert}^{1/2}.$$ | (4.8) |

For an arbitrary fixed $\epsilon >0$, we define the following two sets

${M}_{1}(\epsilon )=\{n\in \mathbb{N}:\Vert {L}_{n}(fg)-({L}_{n}f)({L}_{n}g)\Vert \ge \epsilon \},$ | ||

${M}_{2}(\epsilon )=\{n\in \mathbb{N}:{\Vert {V}_{n}g\Vert}^{1/2}\ge {\displaystyle \frac{\epsilon}{\sqrt{2}\Vert f\Vert}}\}.$ |

On the basis of (4.8), we have ${M}_{1}(\epsilon )\subseteq {M}_{2}(\epsilon )$. Consequently, we can write

$$0\le \delta ({M}_{1}(\epsilon ))=\underset{n\to \mathrm{\infty}}{lim}\frac{1}{n}\sum _{k=1}^{n}{\chi}_{{M}_{1}(\epsilon )}(k)\le \underset{n\to \mathrm{\infty}}{lim}\sum _{k=1}^{n}{\chi}_{{M}_{2}(\epsilon )}(k)=0,$$ |

see (4.7).

This way $0\le \delta ({M}_{1}(\epsilon ))\le 0$ for any $\epsilon >0$, which ensures the relation (4.6). $\mathrm{\square}$

Acknowledgments. The author is thankful to the referee who carefully checked the manuscript. The comments led to several improvements.

## References

- [1] Acu, A.M., Gonska, H., Raşa, I., Grüss- and Ostrowski-type in Approximation Theory, Ukrainian Mathematical Journal, 63(2011), No. 6, 843-864.
- [2] Bardaro, C., Mantellini, I., A Voronovskaja-type theorem for a general class of discrete operators, Rocky Mountain J. Math., 39(3)(2009), 1411-1442.
- [3] Butzer, P.L., Stens, R.L., Sampling theory for not necessarily band-limited functions: An historical overview, SIAM Review, 34(1992), 40-53.
- [4] Chebyshev, P.L., Sur les expressions approximatives des integrales définies par les autres prises entre les même limites, Proc. Math. Soc. Kharkov, 2(1882), 93-98.
- [5] Farcaş, A., An asymptotic formula for Jain’s operators, Stud. Univ. Babeş-Bolyai Math., 57(2012), 511-517.
- [6] Gadjiev, A.D., Orhan, C., Some approximation theorems via statistical convergence, Rocky Mountain J. Math., 32(2002), No. 1, 129-138.
- [7] Gavrea, I. Tachev, G., On the multiplicity of linear operators, Journal of Mathematical Analysis and Applications, 408(2013), 203-208.
- [8] Gonska, H., Raşa, I., Rusu, M.D., C̆ebys̆ev-Grüss-type inequalities revisited, Mathematica Slovaca, 63(2013), No. 5, 1007-1024.
- [9] Gonska, H., Raşa, I., Rusu, M.D., Chebyshev-Grüss-type inequalities via discrete oscillations, Buletinul Academiei de Ştiinţe a Republicii Moldova, Matematica, 1(74), 2014, 63-89.
- [10] Grüss, G., Über das Maximum des absoluten Betrages von $\frac{1}{b-a}{\int}_{a}^{b}f(x)g(x)\mathit{d}x-\frac{1}{{(b-a)}^{2}}{\int}_{a}^{b}f(x)\mathit{d}x{\int}_{a}^{b}g(x)\mathit{d}x$, Mathematische Zeitschrift, 39(1935), 215-226.
- [11] Jain, G.C., Approximation of functions by a new class of linear operators, J. Aust. Math. Soc., 13(3)(1972), 271-276.
- [12] Mirakjan, G.M., Approximation of continuous functions with the aid of polynomials, Dokl. Akad. Nauk. SSSR, 31(1941), 201-205 (in Russian).
- [13] Ries, S., Stens, R.L., Approximation by generalized sampling series, In: Constructive Theory of Functions: Proc. Conf. Varna, Bulgaria, May/June 1984, (Bl. Sendov, P. Petrushev, R. Maalev and S. Tashev Eds.), Publ. House Bulgarian Academy of Sciences, Sofia (1984), 746-756.
- [14] Rusu, M.D., On Grüss-type inequalities for positive linear operators, Stud. Univ. Babeş-Bolyai Math., 56(2011), No. 2, 551-565.
- [15] Szász, O., Generalization of S. Bernstein’s polynomials to the infinite interval, J. Research National Bureau of Standards, 45(1950), 239-245.
- [16] Uchiyama, M., Proofs of Korovkin’s theorems via inequalities, The Amer. Math. Monthly, 110(2003), 334-336.
- [17] Voronovskaja, E., Détermination de la forme asymptotique d’approximation des fonctions par les polynômes de M. Bernstein, C.R. Acad. Sci. URSS (1932), 79-85.

1. Acu, A.M., Gonska, H., Rasa, I.: Grüss- and Ostrowski-type in approximation theory. Ukr. Math. J. 63(6), 843–864 (2011)

2. Bardaro, C., Mantellini, I.: A Voronovskaja-type theorem for a general class of discrete operators. Rocky Mt. J. Math. 39(3), 1411–1442 (2009)

3. Butzer, P.L., Stens, R.L.: Sampling theory for not necessarily band-limited functions: an historical overview. SIAM Rev. 34, 40–53 (1992)

4. Chebyshev, P.L.: Sur les expressions approximatives des integrales définies par les autres prises entre les même limites. Proc. Math. Soc. Kharkov 2, 93–98 (1882)

5. Farcas, A.: An asymptotic formula for Jain’s operators. Stud. Univ. Babe¸s-Bolyai Math. 57, 511–517 (2012)

6. Gadjiev, A.D., Orhan, C.: Some approximation theorems via statistical convergence. Rocky Mt. J. Math. 32(1), 129–138 (2002)

7. Gavrea, I., Tachev, G.: On the multiplicity of linear operators. J. Math. Anal. Appl. 408, 203–208 (2013)

8. Gonska, H., Rasa, I., Rusu,M.D.: C˘ ebys˘ev–Grüss-type inequalities revisited.Math. Slov. 63(5), 1007–1024 (2013)

9. Gonska, H., Rasa, I., Rusu, M.D.: Chebyshev–Grüss-type inequalities via discrete oscillations. Buletinul Academiei de Stiinte a Republicii Moldova, Matematica 1(74), 63–89 (2014)

10. Grüss, G.: Über das Maximum des absoluten Betrages von 1 b−a b a f (x)g(x)dx − 1 (b−a)2 b a f (x)dx b a g(x)dx. Math. Z. 39, 215–226 (1935)

11. Jain, G.C.: Approximation of functions by a new class of linear operators. J. Aust. Math. Soc. 13(3), 271–276 (1972)

12. Mirakjan, G.M.: Approximation of continuous functions with the aid of polynomials. Dokl. Akad. Nauk SSSR 31, 201–205 (1941). (in Russian)

13. Ries, S., Stens, R.L.: Approximation by generalized sampling series. In: Sendov, B., Petrushev, P., Maalev, R., Tashev, S. (eds.) Constructive Theory of Functions: Proc. Conf. Varna, Bulgaria, May/June 1984, pp. 746–756. Publ. House Bulgarian Academy of Sciences, Sofia (1984)

14. Rusu, M.D.: On Grüss-type inequalities for positive linear operators. Stud. Univ. Babes-Bolyai Math. 56(2), 551–565 (2011)

15. Szász, O.: Generalization of S. Bernstein’s polynomials to the infinite interval. J. Res. Natl. Bureau Stand. 45, 239–245 (1950)

16. Uchiyama,M.: Proofs of Korovkin’s theorems via inequalities. Am. Math.Mon. 110, 334–336 (2003)

17. Voronovskaja, E.: Détermination de la forme asymptotique d’approximation des fonctions par les polynômes de M. Bernstein, pp. 79–85. C.R. Acad. Sci. URSS (1932)