Abstract
In this paper, we introduce a new kind of modified Bernstein–Schurer operators based on the concept of (p, q)-integers. We investigate statistical approximation properties, establish a local approximation theorem, give a convergence theorem for the Lipschitz continuous functions, we also obtain a Voronovskaja-type asymptotic formula. Next, we construct the bivariate operators and get some convergence properties. Finally, we give some graphs to illustrate the convergence properties of operators to some functions.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Recently, Mursaleen et al. applied (p, q)-calculus in approximation theory and introduced the (p, q)-analogue of Bernstein operators in [1]. We mention some of their other works as [2–6].
In 2011, Muraru [7] introduced a generalization of the Bernstein–Schurer operators based on q-integers as follows.
for any \(m\in \mathbb {N}\) and \(f\in C([0,1+l])\), l is fixed. In 2013, Ren and Zeng [8] introduced the following modified q-Bernstein–Schurer operators which preserve linear functions.
for \(f\in C([0,1+l]), l\in \mathbb {N}\cup \{0\}\) is fixed, \(n\in \mathbb {N}, 0<q<1\).
In this paper, firstly, we will introduce a generalization of modified Bernstein–Schurer operators based on (p, q)-integers which will be defined in (2) and investigate some approximation properties, secondly, we will construct the bivariate type operators which will be defined in (22) and obtain some convergence properties, finally, we will give some graphics to illustrate the convergence to some functions.
Before introducing the operators, we mention certain definitions based on (p, q)-integers, details can be found in [9–13]. For any fixed real number \(0<q<p\le 1\) and each nonnegative integer k, we denote (p, q)-integers by \([k]_{p,q}\), where
Also (p, q)-factorial and (p, q)-binomial coefficients are defined as follows:
for \(n\ge k\ge 0\). The (p, q)-Binomial expansion is defined by
When \(p=1\), all the definitions of (p, q)-calculus above are reduced to q-calculus.
For \(f\in C(I)\), \(I=[0,1+l]\), \(l\in \mathbb {N}_0\), \(0<q<p\le 1\) and \(n\in \mathbb {N}\), we introduce the (p, q)-analogue of modified Bernstein–Schurer operators as follows:
It is observed that when \(p=1\), \(S_{n,p,q}^{l}(f;x)\) becomes to (1).
2 Auxiliary results
In order to obtain the approximation properties, we need the following lemmas:
Lemma 2.1
For the (p, q)-analogue of modified Bernstein–Schurer operators (2), we have the following equalities
Proof
(3) and (4) are easily obtained from (2) and the definition of (p, q)-integrals. Using (2) and \([k]_{p,q}^2=q^{k-1}[k]_{p,q}+p[k]_{p,q}[k-1]_{p,q}\), we have
Thus, (5) is proved. Next, from (2) and
we get
(9) is proved. Finally, since
and some simple computations, we have
Thus, (7) is proved. \(\square \)
Remark 2.2
Let \(\{p_n\}\) and \(\{q_n\}\) denote sequences such that \(0<q_n<p_n\le 1\). Then by Bohman and Korovkin Theorem and Lemma 2.1, for any \(f\in C(I)\), if \(\lim _{n\rightarrow \infty }p_n=\lim _{n\rightarrow \infty }q_n=1\) and \(\lim _{n\rightarrow \infty }[n]_{p_n,q_n}=\infty \), operators \(S_{n,p,q}^{l}(f;x)\) convergence uniformly to f(x).
Lemma 2.3
Let \(p=\{p_n\}\), \(q=\{q_n\}\), \(0<q_n<p_n\le 1\) be sequences satisfying \(\lim _{n\rightarrow \infty }p_n=\lim _{n\rightarrow \infty }q_n=1\) and \(\lim _{n\rightarrow \infty }q_n^{n}=a\), \(a\in [0,1)\), then we have
where \(\lambda _1, \lambda _2\in (0,1]\) depending on the sequence \(\{q_n\}\).
Proof
From (3) and (4), we get (8). Since \(S_{n,p,q}^{l}\left( (t-x)^2;x\right) =S_{n,p,q}^{l}\left( t^2;x\right) -2xS_{n,p,q}^{l}(t;x)-x^2\) and \(p[n+l-1]_{p,q}=[n+l]_{p,q}-q^{n+l-1}\), we obtain (9). Indeed, we can get (10) easily from (9). Finally, since
Since
and \( 4p^3[n+l-1]_{p,q}[n+l-2]_{p,q}=4[n+l]_{p,q}^2-4q^{n+l-2}[n+l]_{p,q}(2q+p)+4[2]_{p,q}q^{2n+2l-3}, \) \(6p[n+l-1]_{p,q}=6[n+l]_{p,q}-6q^{n+l-1}\), by some computations, we get
From conditions of Lemma 2.3, we have
thus, (13) \(=3a^2x^4. \) From (12), we get
where \(\lambda _1, \lambda _2, \lambda _3\in (0,1]\) depending on the sequence \(\{q_n\}\). \(\square \)
3 Statistical approximation properties
In this section, we present the statistical approximation properties of the operator \(S_{n,p,q}^{l}(f;x)\) by using the Korovkin-type statistical approximation theorem proved in [14].
Let K be a subset of \(\mathbb {N}\), the set of all natural numbers. The density of K is defined by \(\delta (K):=\lim _{n}\frac{1}{n}\sum _{k=1}^{n}\chi _{K}(k)\) provided the limit exists, where \(\chi _{K}\) is the characteristic function of K. A sequence \(x:=\{x_n\}\) is called statistically convergent to a number L if, for every \(\varepsilon > 0\), \(\delta \{n\in \mathbb {N}: |x_n-L|\ge \varepsilon \}=0\). Let \(A:=(a_{jn}), j,n=1,2,\ldots \) be an infinite summability matrix. For a given sequence \(x:=\{x_n\}\), the \(A-\)transform of x, denoted by \(Ax:=((Ax)_j)\), is given by \((Ax)_j=\sum _{k=1}^{\infty }a_{jn}x_n\) provided the series converges for each j. We say that A is regular if \(\lim _{n}(Ax)_j=L\) whenever \(\lim x=L\). Assume that A is a non-negative regular summability matrix. A sequence \(x=\{x_n\}\) is called A-statistically convergent to L provided that for every \(\varepsilon > 0\), \(\lim _{j}\sum _{n: |x_n-L|\ge \varepsilon }a_{jn}=0\). We denote this limit by \(st_A-\lim _n x_n=L\). For \(A=C_1\), the Ces\(\grave{a}\)ro matrix of order one, A-statistical convergence reduces to statistical convergence. It is easy to see that every convergent sequence is statistically convergent but not conversely.
We consider sequences \(p:=\{p_n\}\), \(q:=\{q_n\}\) for \(0<q_n<p_n\le 1\) satisfying
If \(e_i=t^i,\ t\in \mathbb {R^+},\ i=0,1,2,...\) stands for the ith monomial, then we have
Theorem 3.1
Let \(A=(a_{nk})\) be a non-negative regular summability matrix, \(p:=\{p_n\}\) and \(q:=\{q_n\}\) be sequences satisfying (14), then for all \(f\in C(I)\), \(x\in I\), we have
Proof
Obviously
By (5), we have
Now for a given \(\varepsilon > 0\), let us define the following sets:
Then one can see that \(U\subseteq U_1\cup U_2\), so we have
since \(\displaystyle {st_{A}-\lim _{n}p_n=st_{A}-\lim _{n}q_n=1}\) and \(\displaystyle {st_{A}-\lim _{n}[n]_{p_n,q_n}=\infty }\), we have
which imply that the right-hand side of the above inequality is zero, thus we have
Combining (15) and (16), Theorem 3.1 follows from the Korovkin-type statistical approximation theorem established in [14], the proof is completed. \(\square \)
4 Local approximation properties
Let \(f\in C(I)\), endowed with the norm \(||f||=\sup _{x\in I}|f(x)|\). The Peetre’s K-functional is defined by
where \(\delta > 0\) and \(C^2=\left\{ g\in C(I):g',g''\in C(I)\right\} .\) By [15, p. 177, Theorem2.4], there exits an absolute constant \(C > 0\) such that
where
is the second order modulus of smoothness of \(f\in C(I)\).
Now we give a direct local approximation theorem for the operators \(S_{n,p,q}^{l}(f,x)\).
Theorem 4.1
For \(0<q<p\le 1\), \(x\in I\) and \(f\in C(I)\), we have
where C is a positive constant.
Proof
Let \(g\in C^2\). By Taylor’s expansion
and Lemma 2.1, we get
Hence, by (9), we have
On the other hand, by (3), we have
Hence taking infimum on the right hand side over all \(g\in C^2\), we get
By (17), for every \(0<q<p\le 1\), we have
where C is a positive constant. This completes the proof of Theorem 4.1. \(\square \)
Remark 4.2
For any fixed \(x\in I\), \(l\in \mathbb {N}_0\) and \(n\in \mathbb {N}\), let \(p:=\{p_n\}\) and \(q:=\{q_n\}\) are sequences satisfying \(0<p_n<q_n\le 1\), \(\lim _{n}p_n=\lim _{n}q_n=1\) and \(\lim _{n}[n]_{p_n,q_n}=\infty \), we have
These gives us a rate of pointwise convergence of the operators \(S_{n,p_n,q_n}^{l}(f;x)\) to f(x).
Next we study the rate of convergence of the operators \(S_{n,p,q}^{l}(f;x)\) with the help of functions of Lipschitz class \(Lip_M(\alpha )\), where \(M> 0\) and \(0<\alpha \le 1\). A function f belongs to \(Lip_M(\alpha )\) if
We have the following theorem.
Theorem 4.3
Let \(p:=\{p_n\}\) and \(q:=\{q_n\}\) are sequences satisfying \(0<q_n<p_n\le 1\), \(\lim _{n}p_n=\lim _{n}q_n=1\), \(\lim _{n}[n]_{p_n,q_n}=\infty \) and \(f\in Lip_M(\alpha )\), \(0<\alpha \le 1.\) Then we have
Proof
Since \(S_{n,p,q}^{l}(f;x)\) are linear positive operators and \(f\in Lip_M(\alpha )\) (\(0<\alpha \le 1\)), we have
Applying Hölder’s inequality for sums, we obtain
Thus, Theorem 4.3 is proved. \(\square \)
Now, we give a Voronovskaja-type asymptotic formula for \(S_{n,p,q}^{l}(f;x)\).
Theorem 4.4
Let \(p=\{p_n\}\), \(q=\{q_n\}\), \(0<q_n<p_n\le 1\) be sequences satisfying \(\lim _{n\rightarrow \infty }p_n=\lim _{n\rightarrow \infty }q_n=1\), \(\lim _{n}[n]_{p_n,q_n}=\infty \) and \(\lim _{n\rightarrow \infty }q_n^{n}=a\), \(a\in [0,1)\), then we have
where \(\lambda _1\in (0,1]\) depending on the sequence \(\{q_n\}\).
Proof
Let \(x\in [0,1]\) be fixed. By the Taylor formula, we may write
where r(t; x) is the Peano form of the remainder, \(r(t;x)\in C(I)\), using L’Hopital’s rule, we have
Since (8), applying \(S_{n,p,q}^{l}(f;x)\) to (21), we obtain
By the Cauchy–Schwarz inequality, we have
Since \(r^2(x;x)=0\), then it is obtained easily that \(\lim _{n\rightarrow \infty }[n]_{p,q}S_{n,p,q}^{l}\big (r(t;x) (t-x)^2;x\big )=0\) by (11). Thus, from (10), we have
Theorem 4.4 is proved. \(\square \)
5 Construction of bivariate operators and some approximation properties
In this section, we construct a bivariate (p, q)-analogue of modified Bernstein–Schurer operators and get some approximation properties.
For \(f\in C(I_1\times I_2)\), \(I_1\times I_2=[0,1+l_1]\times [0,1+l_2]\), \(l_1, l_2 \in \mathbb {N}_0\), \(x\in I_1\), \(y\in I_2\), \(0<q_{n_1}, q_{n_2}<p_{n_1}, p_{n_2}\le 1\) and \(n_1, n_2\in \mathbb {N}\), the bivariate (p, q)-analogue of modified Bernstein–Schurer operators are defined as follows
Lemma 5.1
Let \(e_{i,j}(x,y)=x^iy^j\), \(i, j\in \mathbb {N}\), \((x,y)\in (I_1\times I_2)\) be the two-dimensional test functions, the bivariate (p, q)-analogue of modified Bernstein–Schurer operators defined in (22) satisfy the following equalities
Remark 5.2
Let \(\{p_{n_1}\}\), \(\{p_{n_2}\}\), \(\{q_{n_1}\}\) and \(\{q_{n_2}\}\) are sequences such that \(0<q_{n_1}, q_{n_2}<p_{n_1}, p_{n_2}\le 1\). Then by [16] and Lemma 5.1, for any \(f\in C(I_1\times I_2)\), if \(\lim _{n_1\rightarrow \infty }p_{n_1}=\lim _{n_1\rightarrow \infty }q_{n_1}=\lim _{n_2\rightarrow \infty }p_{n_2}=\lim _{n_2\rightarrow \infty }q_{n_2}=1\) and \(\lim _{n_1\rightarrow \infty }[n_1]_{p_{n_1},q_{n_1}}=\lim _{n_2\rightarrow \infty }[n_2]_{p_{n_2},q_{n_2}}=\infty \), operators \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) convergence uniformly to f(x, y).
Lemma 5.3
Let \(\{p_{n_1}\}\), \(\{p_{n_2}\}\), \(\{q_{n_1}\}\), \(\{q_{n_2}\}\), \(0<q_{n_1}, q_{n_2}<p_{n_1}, p_{n_2}\le 1\) be sequences satisfying \(\lim _{n_1\rightarrow \infty }p_{n_1}=\lim _{n_1\rightarrow \infty }q_{n_1}=\lim _{n_2\rightarrow \infty }p_{n_2}=\lim _{n_2\rightarrow \infty }q_{n_2}=1\) and \(\lim _{n_1\rightarrow \infty }[n_1]_{p_{n_1},q_{n_1}}=\lim _{n_2\rightarrow \infty }[n_2]_{p_{n_2},q_{n_2}}=\infty \). The following equalities hold
For \(f\in C(I_1\times I_2)\), the complete modulus of continuity for the bivariate case is defined as
where \(\delta _1, \delta _2> 0\). Furthermore, \(\omega (f;\delta _1,\delta _2)\) satisfies the following properties:
The partial moduli of continuity with respect to x and y is defined as
Details of the modulus of continuity for bivariate case can be found in [17].
Now, we give the estimate of the rate of convergence of bivariate (p, q)-analogue of modified Bernstein–Schurer operators defined in (22).
Theorem 5.4
For \(f\in C(I_1\times I_2)\), under the conditions of Lemma 5.3, we have
where \(\delta _{n_1}(x)\) and \(\delta _{n_2}(y)\) are defined in (31) and (32).
Proof
From Lemma 5.1, using the property (ii) above and Cauchy–Schwarz inequality, we easily get
Theorem 5.4 is proved. \(\square \)
Theorem 5.5
For \(f\in C(I_1\times I_2)\), under the conditions of Lemma 5.3, we have
where \(\delta _{n_1}(x)\) and \(\delta _{n_2}(y)\) are defined in (31) and (32).
Proof
Using the definition of partial moduli of continuity above and Cauchy–Schwarz inequality, we have
Theorem 5.5 is proved. \(\square \)
Finally, we study the rate of convergence of \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) by means of functions of Lipschitz class \(Lip_{M}(\alpha _1, \alpha _2)\) if
Theorem 5.6
Let \(f\in Lip_{M}(\alpha _1, \alpha _2)\), under the conditions of Lemma 5.3, we have
where \(\delta _{n_1}(x)\) and \(\delta _{n_2}(y)\) are defined in (31) and (32).
Proof
Since \(f\in Lip_{M}(\alpha _1, \alpha _2)\), we get
using the Hölder’s inequality for last formula, respectively, we obtain
where \(\delta _{n_1}(x)\) and \(\delta _{n_2}(y)\) are defined in (31) and (32). Theorem 5.6 is proved. \(\square \)
6 Graphical analysis
In this section, we give several graphs to show the convergence of \(S^{l}_{n,p,q}(f;x)\) to f(x) and \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) to f(x, y) with different values of parameters.
Let \(f(x) =\sin (\pi x/ 2 )\), for \(n = 50\), \(l=1\) and \(q = 0.9\), the graphs of \(S^{l}_{n,p,q}(f;x)\) with different values of p are shown in Fig. 1. Moreover, let \(f(x) = 1 - \cos (4 e^x)\), for \(n = 50\), \(l=1\) and \(p = 1\), the graphs of \(S^{l}_{n,p,q}(f;x)\) with different values of q are shown in Fig. 2.
Let \( f(x,y) = \sin (x^3 + y^3)\), Fig. 3 shows the graphs of \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) (blue) and f(x, y) (red) for \(n_1=n_2=50\), \(l_1=l_2=1\), \(p_{n_1}=p_{n_2}=1\) and \(q_{n_1}=q_{n_2}=0.6\). Let \(f(x,y) = x^2 \sin (\pi y) + \cos (\pi x) y^2\), in Fig. 4, the values of \(q_{n_1}, q_{n_2}\) are replaced by 0.9, the graphs of \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) (red) and f(x, y) (blue) are shown.
Let \(f(x,y) = x^2 \sin (\pi y) + \cos (\pi x) y^2\), Fig. 5 shows the graphs of \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) (green) and f(x, y) (red) for \(n_1=n_2=50\), \(l_1=l_2=1\), \(p_{n_1}=p_{n_2}=0.999\) and \(q_{n_1}=q_{n_2}=0.9\). Finally, the values of \(p_{n_1}, p_{n_2}\) are substituted for 0.9999, the graphs of \(S_{p_{n_1},p_{n_2},q_{n_1},q_{n_2}}^{n_1,n_2,l_1,l_2}(f;x,y)\) (blue) and f(x, y) (red) are shown in Fig. 6.
References
Mursaleen, M., Ansari, K.J., Khan, A.: On (\(p\), \(q\))-analogue of Bernstein operators. Appl. Math. Comput. 266, 874–882 (2015)
Mursaleen, M., Ansari, K.J., Khan, A.: Some approximation results by (\(p\), \(q\))-analogue of Bernstein–Stancu operators. Appl. Math. Comput. 264, 392–402 (2015)
Mursaleen, M., Ansari, K.J., Khan, A.: Some approximation results for Bernstein–Kantorovich operators based on (\(p\), \(q\))-integers (2015). arXiv:1504.05887v2
Mursaleen, M., Ansari, K.J., Khan, A.: Some approximation results on Bleimann–Butzer–Hahn operators defined by (\(p\), \(q\))-integers (2015). arXiv:1505.00392v1
Mursaleen, M. (2015) Md. Nasiruzzaman, Some approximation properties of bivariate Bleimann–Butzer and Hahn operators based on (\(p\), \(q\))-integers. arXiv:1506.02487v1
Mursaleen, M., Nasiruzzaman, Md, Nurgali, A.: Some approximation results on Bernstein–Schurer operators defined by (\(p\), \(q\))-integers. J. Ineq. Appl. 249, 1–12 (2015)
Muraru, C.-V.: Note on \(q\)-Bernstein–Schurer operators. Stud. Univ. Babeş-Bolyai Math. 56(2), 489–495 (2011)
Ren, M.Y., Zeng, X.M.: On statistical approximation properties of modified \(q\)-Bernstein–Schurer operators. Bull. Korean Math. Soc. 50(4), 1145–1156 (2013)
Hounkonnou, M.N., Désiré, J., Kyemba, B.: \(R\)(\(p\), \(q\))-calculus: differentiation and integration. SUT J. Math. 49, 145–167 (2013)
Jagannathan, R., Rao, K.S.: Two-parameter quantum algebras, twin-basic numbers, and associated generalized hypergeometric series. In: Proceedings of the International Conference on Number Theory and Mathematical Physics, pp. 20–21 (2005)
Katriel, J., Kibler, M.: Normal ordering for deformed boson operators and operator-valued deformed Stirling numbers. J. Phys. A 24, 2683–2691, printed in the UK (1992)
Sadjang, P.N.: On the fundamental theorem of (\(p\), \(q\))-calculus and some (\(p\), \(q\))-Taylor formulas (2015). arXiv:1309.3934v1
Sahai, V., Yadav, S.: Representations of two parameter quantum algebras and \(p\), \(q\)-special functions. J. Math. Anal. Appl. 335, 268–279 (2007)
Duman, O.: A-statistical convergence of sequences of convolution operators. Taiwan. J. Math. 12(2), 523–536 (2008)
DeVore, R.A., Lorentz, G.G.: Constructive Approximation. Springer, Berlin (1993)
Volkov, V.I.: On the convergence of sequences of linear positive operators in the space of continuous functions of two variables (in Russian). Dokl. Akad. Nauk SSSR (NS) 115, 17–19 (1957)
Anastassiou, G.A., Gal, S.G.: Approximation theory: moduli of continuity and global smoothness preservation. Birkhauser, Boston (2000)
Acknowledgments
This work is supported by the National Natural Science Foundation of China (Grant No. 61572020), the China Postdoctoral Science Foundation funded project (Grant no. 2015M582036) and the Startup Project of Doctor Scientific Research and Young Doctor Pre-Research Fund Project of Quanzhou Normal University (Grant No. 2015QBKJ01). I also thank Fujian Provincial Key Laboratory of Data Intensive Computing and Key Laboratory of Intelligent Computing and Information Processing of Fujian Province University.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Cai, QB. On (p, q)-analogue of modified Bernstein–Schurer operators for functions of one and two variables. J. Appl. Math. Comput. 54, 1–21 (2017). https://doi.org/10.1007/s12190-016-0991-1
Received:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12190-016-0991-1
Keywords
- (p, q)-Integers
- Bernstein–Schurer operators
- A-statistical convergence
- Rate of convergence
- Lipschitz continuous functions