Abstract
This study deals with approximation properties by the complex bivariate Balázs-Szabados operators of tensor-product kind. The upper and lower estimates and a Voronovskaja-type theorem of these operators are given. The exact degree of approximation for these operators is obtained.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
The applications on the real and complex approximations properties of the operators have been recently an active subject in the area of the approximation theory (see [7, 10–14]).
Balázs [3] defined the Bernstein type rational functions. He gave an estimate for the order of its convergence and proved an asymptotic approximation theorem and a convergence theorem concerning the derivative of these operators.
In [4] , Balázs and Szabados obtained the best possible estimate under the more restrictive conditions, in which both the weight and the order of convergence would be better than [3]. They applied their results to the approximate certain improper integrals by quadrature sums of positive coefficients based on finite number of equidistant nodes.
Atakut and Ispir [2] defined the bivariate real Bernstein type rational functions of the Bernstein type rational functions given by Balázs [3] and proved the approximation theorems for these functions. In [8], Gupta and Ispir studied on the Bezier variant of generalized Kantorovitch type Balázs operators.
The rational complex Balázs-Szabados operators was defined by Gal [6] as follows:
where \(D_{R}=\left\{ z\in :\left| z\right| <R\right\} \) with \(R>0,\) \(f:D_{R}\cup \left[ R,\infty \right) \rightarrow \) is a function, \(a_{n}=n^{\beta -1},\) \(b_{n}=n^{\beta }\) for \(0<\beta \le 2/3,\) \(n\in \mathbb {N} ,\) \(z\in \mathbb {C} \) with \(z\ne -1/{a_{n}}\). He obtained the uniform convergence of \(R_{n}\big ( f;z\big ) \) to \(f\big ( z\big ) \) on compact disk and proved the upper estimate in approximation of these operators. Also, he obtained the Voronovskaja-type formula and the exact degree of its approximation.
Gal extended some complex univariate operators to the case of several complex variables. To illustrate, he studied the approximation properties of the complex Bernstein polynomials, both tensor-product and non-tensor-product types, the complex Favard-Szász-Mirakjan operators of tensor-product kind without exponential growth conditions on f and the complex Baskakov operators of tensor-product kind ([6], see pp. 155–179).
We consider the following real bivariate Balázs-Szabados operators of the univariate Bernstein type rational functions given by Balázs and Szabados[4]
where \(f:\left[ 0,\infty \right) \times \left[ 0,\infty \right) \rightarrow \mathbb {R} \) is a function, \(a_{n}=n^{\beta -1},\) \(b_{n}=n^{\beta },\) \( a_{m}=m^{\beta -1},\) \(b_{m}=m^{\beta }\) for n, \(m\in \mathbb {N} ,\) \(0<\beta \le 2/3,\) and \(x_{1,}\) \(x_{2}\in \mathbb {R} \) with \(x_{1}\ne -1/{a_{n}}\), \(x_{2}\ne -1/{a_{m}}\),
The real bivariate Balázs-Szabados operators are well defined, linear, and positive. For \(\beta =2/3\), we get the real bivariate Balázs-Szabados operators given by Atakut and Ispir [2].
We define the complex bivariate Balázs-Szabados operators as follows:
where \(f:\left( D_{R_{1}}\cup \left[ R_{1},\infty \right) \right) \times \left( D_{R_{2}}\cup \left[ R_{2},\infty \right) \right) \rightarrow \mathbb {C} \) is a function , \(a_{n}=n^{\beta -1},\) \(b_{n}=n^{\beta },\) \( a_{m}=m^{\beta -1},\) \(b_{m}=m^{\beta },\) for n, \(m\in \mathbb {N} ,\) \(0<\beta \le 2/3,\) and \(z_{1,}\) \(z_{2}\in \mathbb {C} ,\) with \(z_{1}\ne -1/a_{n}\), \(z_{2}\ne -1/{a_{m}}\),
The complex bivariate Balázs-Szabados operators are well defined, linear, and these operators are analytic for all \(n\ge n_{0},\) \(m\ge m_{0}\) , \(\left| z_{1}\right| \le r_{1}<n_{0}^{1-\beta }\) and \(\left| z_{2}\right| \le r_{2}<m_{0}^{1-\beta }\) since \(z_{1}\ne -1/ {a_{n}}\) and \(z_{2}\ne -1/{a_{m}}.\)
Throughout the paper, we denote with \(\left\| f\right\| _{r_{1},r_{2}}=\max \left\{ \left| f\big ( z_{1},z_{2}\big ) \right| :\big ( z_{1},z_{2}\big ) \right. \) \(\left. \in \bar{D} _{r_{1}}\times \bar{D}_{r_{2}}\right\} \) the uniform norm of f in the space of continuous functions on \(\bar{D}_{r_{1}}\times \bar{D}_{r_{2}}\) and with \(\left\| f\right\| _{B\left( \left[ 0,\infty \right) \times \left[ 0,\infty \right) \right) }=\sup \left\{ \left| f\big ( z_{1},z_{2}\big ) \right| :\big ( z_{1},z_{2}\big ) \in \left[ 0,\infty \right) \times \left[ 0,\infty \right) \right\} \) the uniform norm of f in the space of bounded functions on \(\left[ 0,\infty \right) \times \left[ 0,\infty \right) ,\) where \(D_{r}=\left\{ z\in \mathbb {C} :\left| z\right| <r\right\} \) for \(r>0.\)
2 Auxiliary Results
We need following lemmas and theorems in order to prove the main results for the operators (1.1).
Lemma 1
For all \(n,m\in \mathbb {N}, \) we have
where \(\left( e_{i,j}\right) \big ( x_{1},x_{2}\big ) =e_{1}^{i}\left( x_{1}\right) e_{2}^{j}\left( x_{2}\right) \) with \(e_{1}^{i}\left( x_{1}\right) =x_{1}^{i}\) and \(e_{2}^{j}\left( x_{2}\right) =x_{2}^{j}\) for \(i,j=0,1,2.\)
Proof
Using Barbosu technique in [5] and Lemma 2.1 in [3], the lemma can be easily proved, so we will omit the proof of the lemma. \(\square \)
Lemma 2
Let \(f:\left[ 0,\infty \right) \times \left[ 0,\infty \right) \rightarrow \mathbb {R} \) be a continuous function. If
then \(\left\{ R_{n,m}\left( f\right) \right\} \) converges uniformly to f on \(\left[ 0,r_{1}\right] \times \left[ 0,r_{2}\right] \) for \(r_{1},r_{2}>0.\)
Proof
From Lemma 2, taking into account Volkov’s theorem in [15] (also see in [1], p.245),the lemma can be easily proved, so we will omit the proof of the lemma. \(\square \)
Lemma 3
Let \(n_{0},m_{0}\ge 2,\) \(0<\beta \le 2/3\), \(1/2 <r_{1}<R_{1}\le n_{0}^{1-\beta }/2\) and \(1/2 <r_{2}<R_{2}\le m_{0}^{1-\beta }/2\). If \(f:\left( D_{R_{1}}\cup \left[ R_{1},\infty \right) \right) \times \left( D_{R_{2}}\cup \left[ R_{2},\infty \right) \right) \rightarrow \mathbb {C} \) is a uniformly continuous bounded on \(\left[ 0,\infty \right) \times \left[ 0,\infty \right) \) and analytic in \(D_{R_{1}}\times D_{R_{2}}\), then we have the form
for all \(\big ( z_{1},z_{2}\big ) \in D_{r_{1}}\times D_{r_{2}},\) where \( \left( e_{k,j}\right) \big ( z_{1},z_{2}\big ) =e_{1}^{k}\left( z_{1}\right) e_{2}^{j}\left( z_{2}\right) \) with \(e_{1}^{k}\left( z_{1}\right) =z_{1}^{k}\) and \(e_{2}^{j}\left( z_{2}\right) =z_{2}^{j}\) for \( k,j\in \mathbb {N}.\)
Proof
For any \(s,r\in \mathbb {N} ,\) we define
and
From the hypothesis on f, it is clear that each \(f_{s,r}\) is bounded on \( \left[ 0,\infty \right) \times \left[ 0,\infty \right) ,\) which implies that
where \(M_{f_{s,r}}\) is a constant depending on \(f_{s,r},\) so all \( R_{n,m}\left( f_{s,r}\right) \) are well defined for all n, \(m\in \mathbb {N} ,\) \(n\ge n_{0},\) \(m\ge m_{0},\) \(r_{1}<n_{0}^{1-\beta }/2,\) \(r_{2}< m_{0}^{1-\beta }/2\) and \(\big ( z_{1},z_{2}\big ) \in D_{r_{1}}\times D_{r_{2}}\).
Defining
and
It is clear that each \(f_{s,r,k,j}\) is bounded on \(\left[ 0,\infty \right) \times \left[ 0,\infty \right) \) and
From the linearity of \(R_{n,m},\) we have
It suffices to prove that
for any fixed n, \(m\in \mathbb {N} ,\) \(n\ge n_{0},\) \(m\ge m_{0}\), \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}.\) Since
we can write
for \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}.\)
In (2.1), taking limit as \(s,r\rightarrow \infty \) and using \( \lim \limits _{s,r\rightarrow \infty }\left\| f_{s,r}-f\right\| _{r_{1},r_{2}}=0,\) we get the result. \(\square \)
Next lemma and theorem will help to get the convergence result of the operators (1.1).
Lemma 4
Let \(n_{0},m_{0}\ge 2,\) \(0<\beta \le 2/3\), \(1/2 <r_{1}<R_{1}\le n_{0}^{1-\beta }/2\) and \(1/2 <r_{2}<R_{2}\le m_{0}^{1-\beta }/2\). For all \(n\ge n_{0},\) \(m\ge m_{0}\), \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}, \)the following inequality holds
Proof
Using Lemma 1.10.2 in [6] (see p. 141), the lemma is easily proved, so we will omit the proof of the theorem. \(\square \)
3 Main Results
Let us denote with \(A_{C}\left( f\right) \) the space of the all complex valued functions, which are uniformly continuous on \(\left( D_{R_{1}}\cup \left[ R_{1},\infty \right) \right) \times \left( D_{R_{2}}\cup \left[ R_{2},\infty \right) \right) ,\) bounded on \(\left[ 0,\infty \right) \times \left[ 0,\infty \right) \) and analytic in \(D_{R_{1}}\times D_{R_{2}}\), that is \(f\left( z_{1},z_{2}\right) =\sum \nolimits _{k=0}^{\infty }f_{k}\left( z_{2}\right) z_{1}^{k}\)for all \(\left( z_{1},z_{2}\right) \in D_{R_{1}}\times D_{R_{2}}\) with \(f_{k}\left( z_{2}\right) =\sum \nolimits _{j=0}^{\infty }c_{k,j}z_{2}^{j},\) and for which there exist \( M>0,\) \(0<A_{1}<1/2r_{1}\) and \(0<A_{2}<1/2r_{2}\) with \( \left| c_{k,j}\right| \le MA_{1}^{k}A_{2}^{j}/k!j!\) (which implies \(\left| f\left( z_{1},z_{2}\right) \right| \le Me^{A_{1}\left| z_{1}\right| +A_{2}\left| z_{2}\right| }\) for all \(\left( z_{1},z_{2}\right) \in D_{R_{1}}\times D_{R_{2}}).\)
Now, we can give the following convergence result.
Theorem 1
Let \(n_{0},m_{0}\ge 2,\) \(0<\beta \le 2/3\), \(1/2 <r_{1}<R_{1}\le n_{0}^{1-\beta }/2\) and \(1/2 <r_{2}<R_{2}\le m_{0}^{1-\beta }/2\). If \(f\in A_{C}\left( f\right) \), then the sequence of operators \(\left\{ R_{n,m}\left( f\right) \left( z_{1},z_{2}\right) \right\} \) is uniformly convergent to f on \(\bar{D} _{r_{1}}\times \bar{D}_{r_{2}}\) for all \(n\ge n_{0}\) and \(m\ge m_{0}.\)
Proof
From Lemma 3 and Lemma 4, we can write
where the series \(\sum \nolimits _{k=0}^{\infty }\sum \nolimits _{j=0}^{\infty }\big ( 2r_{1}A_{1}\big ) ^{k}\big ( 2r_{2}A_{2}\big ) ^{j}\) is convergent since \(0<A_{1}<1/2r_{1}\) and \(0<A_{2}<1/2r_{2}.\)
On the other hand, from Lemma 2, we know that
for all \(\big ( x_{1},x_{2}\big ) \in \left[ 0,r_{1}\right] \times \left[ 0,r_{2}\right] .\)
Using (3.1) and (3.2) and taking into account the Vitali’s theorem [9] (see p.112, Theorem 3.2.10), it follows that \(\left\{ R_{n,m}\left( f\right) \big ( z_{1},z_{2}\big ) \right\} \) uniformly converges to f on \(\bar{D}_{r_{1}}\times \bar{D}_{r_{2}}\) for all \(n\ge n_{0}\) and \(m\ge m_{0}.\)
We have the following upper estimate.
Theorem 2
Let \(n_{0},m_{0}\ge 2,\) \(0<\beta \le 2/3\), \(1/2 <r_{1}<R_{1}\le n_{0}^{1-\beta }/2\) and \(1/2 <r_{2}<R_{2}\le m_{0}^{1-\beta }/2\). If \(f\in A_{C}\left( f\right) \), then for all \(n\ge \max \left\{ n_{0},1/r_{1}^{1/\beta }\right\} ,\) \(m\ge \max \left\{ m_{0},1/r_{2}^{1/\beta }\right\} ,\) \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}\) the following inequality holds
where
and also the series \(\sum \nolimits _{k=0}^{\infty }\big ( 2r_{1}A_{1}\big ) ^{k},\) \(\sum \nolimits _{k=1}^{\infty }k\big ( 2r_{1}A_{1}\big ) ^{k-1}\) and \( \sum \nolimits _{j=1}^{\infty }j\big ( 2r_{2}A_{2}\big ) ^{j-1}\) are convergent.
Proof
Using Lemma 3, we can write
Taking into account Lemma 1.10.2 and the estimate given in the proof of Theorem 1.10.5 in [6] (see p. 141 pp 144-145), for all \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2},\) we obtain
Applying (3.4) in(3.3), we get
which complete the prove. \(\square \)
In what follows a Voronovskaja-type result for the operators (1.1) is presented. It will be the product of the parametric extensions generated by Voronovskaja’s formula in univariate case in Theorem 1.10.6 in [6] (see p. 145).
Theorem 3
Let \(n_{0},m_{0}\ge 2,\) \(0<\beta \le 2/3\), \(1/2 <r_{1}<R_{1}\le n_{0}^{1-\beta }/2\) and \(1/2 <r_{2}<R_{2}\le m_{0}^{1-\beta }/2\). If \(f\in A_{C}\left( f\right) \), then for all \(n\ge \max \left\{ n_{0},1/r_{1}^{1/\beta }\right\} ,\) \(m\ge \max \left\{ m_{0},1/r_{2}^{1/\beta }\right\} ,\) \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}\) the following inequality holds
where \(C^{3}\left( f\right) =1/2\max \left\{ C_{r_{1},r_{2}}^{r_{1}^{*}}\left( f\right) ,C_{r_{1},r_{2}}^{r_{2}^{*}}\left( f\right) \right\} \) with
for \(1/2<r_{1}<r_{1}^{*}<n_{0}^{1-\beta }/2\) and \( 1/2<r_{2}<r_{2}^{*}\le m_{0}^{1-\beta }/2.\)
Proof
For \(f\left( z_{1},z_{2}\right) ,\) we define the parametric extensions of the Voronovskaja’s formula by
and
Their product (composition) gives
After simple calculation, it is obtained the commutativity property
From the analyticity of f, since the all partial derivatives of f are analytic in \(D_{R_{1}}\times D_{R_{2}},\) using Lemma 3, we can write
Applying now \(R_{m}\) to (3.6) with respect to \(z_{2}\), we obtain
In (3.7), passing now to absolute value for \(\left| z_{1}\right| \le r_{1}\) and \(\left| z_{2}\right| \le r_{2}\) and taking into account the Lemma 1.10.2 and the estimate given in the proof of Theorem 1.10.6 in [6] (see p. 141 and pp. 145–146), it follows
where \(r_{1}<r_{1}^{*}<R<n_{0}^{1-\beta }/2\) with \(r_{1}^{*}< 1/2A_{1}\) and \(C_{r_{1,}r_{1}^{*}}\) is a constant depending on \( r_{1}\) and \(r_{1}^{*}.\)
Similarly, using Lemma 3, we have
Using
we can write
which implies
And also, using
we can write
which implies
If we estimate \(\left| z_{1}T_{n}\left( f\right) \left( z_{1},z_{2}\right) \circ z_{2}T_{m}\left( f\right) \left( z_{1},z_{2}\right) \right| ,\) then by reason of symmetry we get a similar order of approximation, simply interchanging above the places of n with m and \( r_{1}\) with \(r_{2}.\)
In conclusion, using the commutativity property given in (3.5), we reach the result. \(\square \)
Let us denote with \(A_{C}^{\left( 2\right) }\left( f\right) \) the space of the all complex valued functions that they and their first and second partial derivatives are uniformly continuous on \(\left( D_{R_{1}}\cup \left[ R_{1},\infty \right) \right) \times \left( D_{R_{2}}\cup \left[ R_{2},\infty \right) \right) ,\) bounded on \(\left[ 0,\infty \right) \times \left[ 0,\infty \right) \) and analytic in \(D_{R_{1}}\times D_{R_{2}}\) and there exist \(M>0,\) \(0<A_{1}<1/2r_{1},\) \(0<A_{2}<1/2r_{2}\) with \( \left| c_{k,j}\right| \le MA_{1}^{k}A_{2}^{j}/k!j! (\) which implies \(\left| f\left( z_{1},z_{2}\right) \right| \le Me^{A_{1}\left| z_{1}\right| +A_{2}\left| z_{2}\right| }\) for all \(\left( z_{1},z_{2}\right) \in D_{R_{1}}\times D_{R_{2}}).\)
Theorem 2 and Theorem 3 will be used to find the exact degree in approximation of \(R_{n,n}\left( f\right) .\) In this sense we have the following lower estimate.
Theorem 4
Let \(n_{0}\ge 2,\) \(0<\beta <1/2\), \(1/2<r_{1}<R_{1}\le n_{0}^{1-\beta }/2\), and \(1/2<r_{2}<R_{2}\le n_{0}^{1-\beta }/2\). If \(f\in A_{C}^{\left( 2\right) }\left( f\right) \ \) and f is not a solution of the complex partial differential equation
then for all \(n\ge n_{0}\) we have
Proof
We can write
where
with
and
Under the conditions of theorem, since \(\lim \limits _{n\rightarrow \infty }a_{n}=0\), \(\lim \limits _{n\rightarrow \infty }1/b_{n}= 0\) and \(\lim \limits _{n\rightarrow \infty }a_{n}b_{n}= 0\) for \(0<\beta <1/2\), considering Theorem 1.10.5 and Theorem 1.10.6 in [6] (see p. 145), it is clear that
From Theorem 3, we obtain
Using \(\lim \limits _{n\rightarrow \infty }a_{n}b_{n}= 0 \) for \(0<\beta <1/2\) and \(1 /1+a_{n}\left| z_{1}\right| \ge 2/3,\) we get
Similarly, it follows
From (3.15) and (3.16), we can write
In (3.13), taking into account the inequalities
and (3.17), it follows
for all \(n\ge n_{0}\) with \(n_{0}\) depending only \(f,r_{1}\) and \(r_{2.}\) We used here that by hypothesis we have \(\left\| K\left( f\right) \right\| _{r_{1},r_{2}}>0.\) \(\square \)
Combining Theorem 3 with Theorem 4, we immediately obtain the following result giving the exact degree of the operators (1.1).
Corollary 1
Suppose that the hypothesis in the statement of Theorem 6 holds. If Taylor series of f contains at least one term of the form \(c_{k+1,0}z_{1}^{k}\) with \(c_{k+1,0}\ne 0\) and \(k=1,2,\ldots \) or of the form \(c_{0,j+1}z_{2}^{j}\) with \(c_{0,j+1}\ne 0\) and \(j=1,2,\ldots \), then for all \(n\ge n_{0}\) we have
Proof
It suffices to prove that under the hypothesis on f, it cannot be a solution of the complex partial differential equation
Indeed, suppose the contrary. Since simple calculation gives
by equaling with zero and by the identification of coefficients, from the terms under the first and fifth sign \(\sum ,\) we immediately get that \( c_{k+1,0}=c_{0,j+1}=0,\) for all \(k=1,2,\ldots \) and \(j=1,2,\ldots ,\) which contradicts the hypothesis on f. Therefore the hypothesis and the lower estimate in Theorem 4 satisfy, which complete the proof. \(\square \)
References
Altomare, F., Campiti, M.: Korovkin-Type Approximation Theory and Its Applications. Walter de Gruyter, Berlin (1994)
Atakut, Ç., İspir, N.: On Bernstein type rational functions of two variables. Math. Slovacka 54(3), 291–301 (2004)
Balázs, K.: Approximation by Bernstein type rational function. Acta Math. Acad. Sci. Hungar 26, 123–134 (1975)
Balázs, K., Szabados, J.: Approximation by Bernstein type rational function II. Acta Math. Acad. Sci. Hungar 40(3–4), 331–337 (1982)
Barbosu, D.: Some generalized bivariate Bernstein operators. Math. Notes (Miskolc) 1, 3–10 (2000)
Gal, S.G.: Approximation by Complex Bernstein and Convolution Type Operators. World Scientific Publishing Co. Pte.Ltd., New Jersey (2009)
Gal, S.G., Gupta, V., Mahmudov, N.I.: Approximation by a complex q-Durmeyer type operator. Ann. Univ. Ferrara 58, 65–87 (2012)
Gupta, V., Ispir, N.: On the Bezier variant of generalized Kantorovitch type Balazs operators. Appl. Math. Lett. 18(9), 1053–1061 (2005)
Kohr, G., Mocanu, P.T.: Special Chapters of Complex Analysis. University Press, Cluj-Napuca (2005). (in Romania)
Mahmudov, N.I.: Approximation by Bernstein-Durmeyer-type Polynomials in compact disks. Appl. Math. Lett. 24, 1231–1238 (2011)
Mahmudov, N.I.: Approximation by genuine q-Bernstein-Durmeyer operators in compact disks. Hacet J. Math. Stat. 40(1), 77–89 (2011)
Mursaleen, M., Khan, A.: Statistical approximation properties of modified q-Stancu-Beta operators. Bull. Malays. Math. Sci. Soc. (2) 36(3), 683–690 (2013)
Ispir, N., Yıldız Özkan, E.: Approximation properties of complex q-Balazs-Szabados operators in compact disks. J. İnequal. Appl. 2013, 361 (2013). doi:10.1186/1029-242X-2013-361
Agarwal, R.P., Gupta, V.: On q-analogue of a complex summation-integral type operators in compact disks. J. İnequal. Appl. 2012, 111 (2012)
Volkov, V.I.: On the convergence of sequences of linear positive operators in the space of continuous functions of two variables. (Russian) Dokl. Akad. Nauk. SSSR (N.S.) 115 (1957)
Acknowledgments
The author is grateful to the editor and the reviewers for making valuable suggestions, leading to a better presentation of the work.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Lee See Keong.
Rights and permissions
About this article
Cite this article
Özkan, E.Y. Approximation by Complex Bivariate Balázs-Szabados Operators. Bull. Malays. Math. Sci. Soc. 39, 1–16 (2016). https://doi.org/10.1007/s40840-015-0159-4
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40840-015-0159-4