
This paper introduced a novel class of bivariate λ-Bernstein operators defined on triangular domain, denoted as Bλ1,λ2m(f;x,y). These operators leverage a new class of bivariate Bézier basis functions defined on triangular domain with shape parameters λ1 and λ2. A Korovkin-type approximation theorem for Bλ1,λ2m(f;x,y) was established, with the convergence rate being characterized by both the complete and partial moduli of continuity. Additionally, a local approximation theorem and a Voronovskaja-type asymptotic formula were derived for Bλ1,λ2m(f;x,y). Finally, the convergence of Bλ1,λ2m(f;x,y) to f(x,y) was illustrated through graphical representations and numerical examples, highlighting instances where they surpass the performance of standard bivariate Bernstein operators defined on triangular domain, Bm(f;x,y).
Citation: Guorong Zhou, Qing-Bo Cai. Bivariate λ-Bernstein operators on triangular domain[J]. AIMS Mathematics, 2024, 9(6): 14405-14424. doi: 10.3934/math.2024700
[1] | Mohammad Ayman-Mursaleen, Md. Nasiruzzaman, Nadeem Rao, Mohammad Dilshad, Kottakkaran Sooppy Nisar . Approximation by the modified λ-Bernstein-polynomial in terms of basis function. AIMS Mathematics, 2024, 9(2): 4409-4426. doi: 10.3934/math.2024217 |
[2] | Mustafa Kara . Approximation properties of the new type generalized Bernstein-Kantorovich operators. AIMS Mathematics, 2022, 7(3): 3826-3844. doi: 10.3934/math.2022212 |
[3] | Qing-Bo Cai, Melek Sofyalıoğlu, Kadir Kanat, Bayram Çekim . Some approximation results for the new modification of Bernstein-Beta operators. AIMS Mathematics, 2022, 7(2): 1831-1844. doi: 10.3934/math.2022105 |
[4] | Purshottam Narain Agrawal, Behar Baxhaku, Abhishek Kumar . Approximation properties of generalized Baskakov operators. AIMS Mathematics, 2021, 6(7): 6986-7016. doi: 10.3934/math.2021410 |
[5] | Zhongbin Zheng, Jinwu Fang, Wentao Cheng, Zhidong Guo, Xiaoling Zhou . Approximation properties of modified (p, q)-Szász-Mirakyan-Kantorovich operators. AIMS Mathematics, 2020, 5(5): 4959-4973. doi: 10.3934/math.2020317 |
[6] | B. B. Jena, S. K. Paikray, S. A. Mohiuddine, Vishnu Narayan Mishra . Relatively equi-statistical convergence via deferred Nörlund mean based on difference operator of fractional-order and related approximation theorems. AIMS Mathematics, 2020, 5(1): 650-672. doi: 10.3934/math.2020044 |
[7] | Sima Karamseraji, Shokrollah Ziari, Reza Ezzati . Approximate solution of nonlinear fuzzy Fredholm integral equations using bivariate Bernstein polynomials with error estimation. AIMS Mathematics, 2022, 7(4): 7234-7256. doi: 10.3934/math.2022404 |
[8] | Nazmiye Gonul Bilgin, Yusuf Kaya, Melis Eren . Security of image transfer and innovative results for (p,q)-Bernstein-Schurer operators. AIMS Mathematics, 2024, 9(9): 23812-23836. doi: 10.3934/math.20241157 |
[9] | Ruifeng Wu . Bivariate multiquadric quasi-interpolation operators of Lidstone type. AIMS Mathematics, 2023, 8(9): 20914-20932. doi: 10.3934/math.20231065 |
[10] | Ruifeng Wu . A kind of even order Bernoulli-type operator with bivariate Shepard. AIMS Mathematics, 2023, 8(7): 15299-15316. doi: 10.3934/math.2023782 |
This paper introduced a novel class of bivariate λ-Bernstein operators defined on triangular domain, denoted as Bλ1,λ2m(f;x,y). These operators leverage a new class of bivariate Bézier basis functions defined on triangular domain with shape parameters λ1 and λ2. A Korovkin-type approximation theorem for Bλ1,λ2m(f;x,y) was established, with the convergence rate being characterized by both the complete and partial moduli of continuity. Additionally, a local approximation theorem and a Voronovskaja-type asymptotic formula were derived for Bλ1,λ2m(f;x,y). Finally, the convergence of Bλ1,λ2m(f;x,y) to f(x,y) was illustrated through graphical representations and numerical examples, highlighting instances where they surpass the performance of standard bivariate Bernstein operators defined on triangular domain, Bm(f;x,y).
Bernstein polynomials and their associated operators have been a cornerstone of the approximation theory, attracting significant research attention in recent years. Researchers have explored diverse generalizations and extensions of these operators, establishing their approximation properties [1,2,3,4,5,6].
In 1963, Stancu [7] introduced a collection of bivariate Bernstein operators denoted as Bm(f;x,y), designed for functions f:S→R:
Bm(f;x,y)=m∑i=0m−i∑j=0pm,i,j(x,y)f(im,jm),(x,y)∈S, | (1.1) |
where S represents the triangular domain {(x,y)|x,y≥0,x+y≤1}, and pm,i,j(x,y) are the associated bivariate Bernstein basis functions:
pm,i,j(x,y)=(mi)(m−ij)xiyj(1−x−y)m−i−j. | (1.2) |
Since then, researchers have actively explored various generalizations of Bernstein-type operators to the triangular domain.
In 2018, Cai et al. [8] introduced a novel generalization of Bernstein operators termed λ-Bernstein operators, defined as:
Bn,λ(f;x)=n∑k=0˜bn,k(λ;x)f(kn),x∈[0,1] and f∈C[0,1], |
where ˜bn,k(λ;x) represent Bézier basis functions with a shape parameter λ∈[−1,1] [9]:
{˜bn,0(λ;x)=bn,0(x)−λn+1bn+1,1(x),˜bn,k(λ;x)=bn,k(x)+λ[n−2k+1n2−1bn+1,k(x)−n−2k−1n2−1bn+1,k+1(x)],(1≤k≤n−1),˜bn,n(λ;x)=bn,n(x)−λn+1bn+1,n(x). |
where bn,k(x)=(nk)xk(1−x)n−k, (k=0,1,…,n) are the Bernstein basis functions. This generalization reduces to the classical Bernstein operators, Bn(f;x), when λ=0. Recent years have witnessed a surge in interest surrounding λ-Bernstein operators Bn,λ(f;x) [10,11,12,13,14,15,16], attracting numerous researchers to explore their properties and applications.
The main purpose of this paper is to construct and study the approximation properties of bivariate λ-Bernstein operators on the triangular domain S. The paper is organized as follows: In Section 2, we present a novel class of bivariate Bézier basis functions on triangular domain S with shape parameters λ1 and λ2, denoted by p(λ1,λ2)m,i,j(x). These functions serve as the foundation for constructing a new kind of bivariate λ-Bernstein operators, Bλ1,λ2m(f;x,y). In Section 3, we delve into the fundamental properties of p(λ1,λ2)m,i,j(x), including nonnegativity, partition of unity, and endpoint properties. Furthermore, we present estimations for the moments and central moments of Bλ1,λ2m(f;x,y). Section 4 establishes theoretical results pertaining to the approximation capabilities of Bλ1,λ2m(f;x,y), including the Korovkin-type approximation theorem, the rate of convergence of Bλ1,λ2m(f;x,y), the local approximation theorem, and a Voronovskaja type theorem. In Section 5, we provide empirical evidence for the convergence behavior through graphical representations and numerical examples, showcasing how Bλ1,λ2m(f;x,y) approaches f(x,y) under various parameter configurations. Finally, we summarize the key findings of this paper in Section 6.
Building upon the construction of Bézier basis functions with a shape parameter λ presented in [9] and the bivariate Bernstein basis functions on triangular domain defined in (1.2), we introduce a novel class of bivariate Bézier basis functions on the triangular domain S with shape parameters λ1 and λ2 as follows:
p(λ1,λ2)m,i,j(x,y)=pm,i,j(x,y)+Λ1i,jpm+1,i,j(x,y)−Λ1i+1,jpm+1,i+1,j(x,y)+Λ2i,jpm+1,i,j(x,y)−Λ2i,j+1pm+1,i,j+1(x,y),(x,y)∈S, | (2.1) |
where Λ1i,j and Λ2i,j are given by
{Λ10,j=Λ2i,0=0,0≤i,j≤m,Λ1i,j=λ1m−2i+1m2−1,1≤i≤m, 0≤j≤m−i,Λ2i,j=λ2m−2j+1m2−1,1≤j≤m, 0≤i≤m−j,Λ1i+1,j=Λ2i,j+1=0,i+j=m, | (2.2) |
with |λ1|≤12, |λ2|≤12. Motivated by the newly introduced class of bivariate Bézier basis functions p(λ1,λ2)m,i,j(x,y), we formally define the associated bivariate λ-Bernstein operators on the triangular domain.
Definition 2.1. Let f∈C[S] and |λ1|≤12, |λ2|≤12. The bivariate λ-Bernstein operators on the triangular domain S are defined as:
Bλ1,λ2m(f;x,y)=m∑i=0m−i∑j=0p(λ1,λ2)m,i,j(x,y)f(im,jm),(x,y)∈S, | (2.3) |
where p(λ1,λ2)m,i,j(x,y) are the new bivariate Bézier basis functions on triangular domain defined in (2.1).
Remark 2.1. It is noteworthy that when λ1=λ2=0, the operators Bλ1,λ2m(f;x,y) coincide with the classical bivariate Bernstein operators on triangular domain Bm(f;x,y) defined in (1.1).
Prior to delving into the properties of p(λ1,λ2)m,i,j(x,y), it is beneficial to recall the key characteristics of their classical counterparts, the bivariate Bernstein basis functions pm,i,j(x,y).
● Nonnegativity:
pm,i,j(x,y)≥0, (x,y)∈S, 0≤i+j≤m. |
● Partition of Unity:
m∑i=0m−i∑j=0pm,i,j(x,y)≡1. |
● Endpoint properties:
pm,i,j(0,0)={1,i=j=0,0else,pm,i,j(1,0)={1,i=m,j=0,0else,pm,i,j(0,1)={1,i=0,j=m,0else. |
We can demonstrate that p(λ1,λ2)m,i,j(x,y) inherit these fundamental properties.
Theorem 3.1. The bivariate Bézier basis function on triangular domain with shape parameters λ1 and λ2, p(λ1,λ2)m,i,j(x,y), as defined in (2.1), have the following properties:
● Nonnegativity:
p(λ1,λ2)m,i,j(x,y)≥0, (x,y)∈S, 0≤i+j≤m. |
● Partition of Unity:
m∑i=0m−i∑j=0p(λ1,λ2)m,i,j(x,y)≡1. |
● Endpoint properties:
p(λ1,λ2)m,i,j(0,0)={1,i=j=0,0,else,p(λ1,λ2)m,i,j(1,0)={1,i=m,j=0,0,else,p(λ1,λ2)m,i,j(0,1)={1,i=0,j=m,0,else. |
Proof. (Nonnegativity) Within the triangular domain S, we can leverage (1.2) and (2.1) to express p(λ1,λ2)m,i,j(x,y) as:
p(λ1,λ2)m,i,j(x,y)=pm,i,j(x,y)[1+Λ1i,jm+1m+1−i−j(1−x−y)−Λ1i+1,jm+1i+1x+Λ2i,jm+1m+1−i−j(1−x−y)−Λ2i,j+1m+1j+1y]. |
For i≥1, j≥1, and i+j≤m−1 cases, since |λ1|≤12,|λ2|≤12, we have
p(λ1,λ2)m,i,j(x,y)=pm,i,j(x,y)[1+λ1m−2i+1(m−1)(m+1−i−j)(1−x−y)−λ1m−2i−1(m−1)(i+1)x+λ2m−2j+1(m−1)(m+1−i−j)(1−x−y)−λ2m−2j−1(m−1)(j+1)y]≥pm,i,j(x,y)[1−1−x−y2−x2−1−x−y2−y2]=pm,i,j(x,y)x+y2≥0. |
For i=0 or j=0, and i+j≤m−1 cases, we have
p(λ1,λ2)m,0,0(x,y)=pm,0,0(x,y)(1−λ1x−λ2y)≥pm,0,0(x,y)[1−12(x+y)]≥0;p(λ1,λ2)m,i,0(x,y)≥pm,i,0(x,y)[1−12(1−x−y)−12x−12y]=12pm,i,0(x,y)≥0;p(λ1,λ2)m,0,j(x,y)≥pm,0,j(x,y)[1−12x−12(1−x−y)−12y]=12pm,0,j(x,y)≥0. |
For i+j=m cases, we have
p(λ1,λ2)m,0,m(x,y)≥pm,0,m(x,y)[1−1−x−y2]=pm,0,m(x,y)1+x+y2≥0,p(λ1,λ2)m,m,0(x,y)≥pm,m,0(x,y)[1−1−x−y2]=pm,m,0(x,y)1+x+y2≥0,p(λ1,λ2)m,i,m−i(x,y)≥pm,i,m−i(x,y)[1−1−x−y2−1−x−y2]=pm,i,m−i(x,y)(x+y)≥0, i≠0,m. |
Therefore, p(λ1,λ2)m,i,j(x,y)≥0 for all (x,y)∈S and 0≤i+j≤m, proving the nonnegativity.
(Partition of Unity)
m∑i=0m−i∑j=0p(λ1,λ2)m,i,j(x,y)=m∑i=0m−i∑j=0pm,i,j(x,y)+m∑i=1m−i∑j=0Λ1i,jpm+1,i,j(x,y)−m−1∑i=0m−i−1∑j=0Λ1i+1,jpm+1,i+1,j(x,y)+m−1∑i=0m−i∑j=1Λ2i,jpm+1,i,j(x,y)−m−1∑i=0m−i−1∑j=0Λ2i,j+1pm+1,i,j+1(x,y). |
We can rearrange with common indices, letting i′=i+1, j′=j+1.
m∑i=0m−i∑j=0p(λ1,λ2)m,i,j(x,y)=1+m∑i=1m−i∑j=0Λ1i,jpm+1,i,j(x,y)−m∑i′=1m−i′∑j=0Λ1i′,jpm+1,i′,j(x,y)+m−1∑i=0m−i∑j=1Λ2i,jpm+1,i,j(x,y)−m−1∑i=0m−i∑j′=1Λ2i,j′pm+1,i,j′(x,y)=1. |
Therefore, we obtain the partition of unity property of p(λ1,λ2)m,i,j(x,y).
(Endpoint properties) The endpoint properties of p(λ1,λ2)m,i,j(x,y) can be directly inherited from the corresponding properties of pm,i,j(x,y) due to their construction and relationship.
Figure 1 presents the graphs of p(λ1,λ2)m,i,j(x,y) for m=2, λ1=0.5, and λ2=−0.5. These visualizations effectively illustrate the nonnegativity and endpoint properties of p(λ1,λ2)m,i,j(x,y).
Remark 3.1. Building upon the nonnegativity and partition of unity of p(λ1,λ2)m,i,j(x,y), we can further deduce that the associated operators Bλ1,λ2m(f;x,y) possess the characteristic of linear positivity.
Lemma 3.1. Let f(s,t)=sktl=ekl,(s,t)∈S, where k,l∈N and 0≤k+l≤4, then for the bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y), we have the following equalities:
Bλ1,λ2m(e00;x,y)=1, | (3.1) |
Bλ1,λ2m(e10;x,y)=x+λ11−2x−(1−x)m+1−(x+y)m+1+ym+1+2x(x+y)m+1m(m−1), | (3.2) |
Bλ1,λ2m(e01;x,y)=y+λ21−2y−(1−y)m+1−(x+y)m+1+xm+1+2y(x+y)m+1m(m−1), | (3.3) |
Bλ1,λ2m(e20;x,y)=x2+x(1−x)m+λ1[2x−4x2−2x(x+y)m+4x2(x+y)m−1m(m−1)−1−(1−x)m+1−(x+y)m+1+ym+1m2(m−1)], | (3.4) |
Bλ1,λ2m(e02;x,y)=y2+y(1−y)m+λ2[2y−4y2−2y(x+y)m+4y2(x+y)m−1m(m−1)−1−(1−y)m+1−(x+y)m+1+xm+1m2(m−1)], | (3.5) |
Bλ1,λ2m(e11;x,y)=xy−xym+λ1[y−2xy−y(1−x)m−y(x+y)m+ym+1+2xy(x+y)m−1m(m−1)+y−y(1−x)m−y(x+y)m+ym+1m2(m−1)]+λ2[x−2xy−x(1−y)m−x(x+y)m+xm+1+2xy(x+y)m−1m(m−1)+x−x(1−y)m−x(x+y)m+xm+1m2(m−1)] | (3.6) |
Bλ1,λ2m(e30;x,y)=x3+3x2−3x3m+x−3x2+2x3m2+λ1[3x2−6x3−3x2(x+y)m−1+6x3(x+y)m−2m2−6x2−6x2(x+y)m−1m2(m−1)+1−2x−(1−x)m+1−(x+y)m+1+ym+1+2x(x+y)mm3(m−1)], | (3.7) |
Bλ1,λ2m(e03;x,y)=y3+3y2−3y3m+y−3y2+2y3m2+λ2[3y2−6y3−3y2(x+y)m−1+6y3(x+y)m−2m2−6y2−6y2(x+y)m−1m2(m−1)+1−2y−(1−y)m+1−(x+y)m+1+xm+1+2y(x+y)mm3(m−1)], | (3.8) |
Bλ1,λ2m(e40;x,y)=x4+6x3−6x4m+7x2−18x3+11x4m2+x−7x2+12x3−6x4m3++λ1[4x3−8x4−4x3(x+y)m−2+8x4(x+y)m−3m2+6x2−32x3+16x4−6x2(x+y)m−1+32x3(x+y)m−2−16x4(x+y)m−3m3+2x−16x2−2x(x+y)m+16x2(x+y)m−1m3(m−1)−1−(1−x)m+1−(x+y)m+1+ym+1m4(m−1)], | (3.9) |
Bλ1,λ2m(e04;x,y)=y4+6y3−6y4m+7y2−18y3+11y4m2+y−7y2+12y3−6y4m3++λ2[4y3−8y4−4y3(x+y)m−2+8y4(x+y)m−3m2+6y2−32y3+16y4−6y2(x+y)m−1+32y3(x+y)m−2−16y4(x+y)m−3m3+2y−16y2−2y(x+y)m+16y2(x+y)m−1m3(m−1)−1−(1−y)m+1−(x+y)m+1+xm+1m4(m−1)]. | (3.10) |
The conclusions of Lemma 3.1 can be derived via more intuitive calculations, hence we omit the proof process here. Based on Lemma 3.1, we present the following Lemma 3.2, established through direct calculations.
Lemma 3.2. Let f(s,t)=(s−x)k(t−y)l=hkl, where k,l∈N and 1≤k+l≤2, then for the bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y), we have the following equalities:
Bλ1,λ2m(h10;x,y)=λ11−2x−(1−x)m+1−(x+y)m+1+ym+1+2x(x+y)m+1m(m−1), | (3.11) |
Bλ1,λ2m(h01;x,y)=λ21−2y−(1−y)m+1−(x+y)m+1+xm+1+2y(x+y)m+1m(m−1), | (3.12) |
Bλ1,λ2m(h20;x,y)=x(1−x)m+λ1[1m(m−1)[4x2(x+y)m−1−(2x+4x2)(x+y)m+2x(1−x)m+1+2x(x+y)m+1−2xym+1]−1−(1−x)m+1−(x+y)m+1+ym+1m2(m−1)], | (3.13) |
Bλ1,λ2m(h02;x,y)=y(1−y)m+λ2[1m(m−1)[4y2(x+y)m−1−(2y+4y2)(x+y)m+2y(1−y)m+1+2y(x+y)m+1−2yxm+1]−1−(1−y)m+1−(x+y)m+1+xm+1m2(m−1)], | (3.14) |
Bλ1,λ2m(h11;x,y)=−xym+λ1[1m(m−1)[(1−y)ym+1+2xy(1−x−y)(x+y)m−1−xy(1−x)m−y(1−x−y)(x+y)m]+1m2(m−1)[y−y(1−x)m−y(x+y)m+ym+1]]+λ2[1m(m−1)[(1−x)xm+1+2xy(1−x−y)(x+y)m−1−xy(1−y)m−x(1−x−y)(x+y)m]+1m2(m−1)[x−x(1−y)m−x(x+y)m+xm+1]]. | (3.15) |
Corollary 3.1. Consider a fixed point (x,y)∈S. Under |λ1|≤12 and |λ2|≤12, we have
limm→∞mBλ1,λ2m((s−x);x,y)=0, | (3.16) |
limm→∞mBλ1,λ2m((t−y);x,y)=0, | (3.17) |
limm→∞mBλ1,λ2m((s−x)2;x,y)=x(1−x), | (3.18) |
limm→∞mBλ1,λ2m((t−y)2;x,y)=y(1−y), | (3.19) |
limm→∞mBλ1,λ2m((s−x)(t−y);x,y)=−xy, | (3.20) |
limm→∞m2Bλ1,λ2m((s−x)4;x,y)=3x2−6x3+3x4, | (3.21) |
limm→∞m2Bλ1,λ2m((t−y)4;x,y)=3y2−6y3+3y4. | (3.22) |
Since
1−(x+y)m+1≥0,1−(1−x)m+1−(x+y)m+1+ym+1≥0,1−(1−y)m+1−(x+y)m+1+xm+1≥0, |
when (x,y)∈S, combining (3.11), (3.13), and |λ1|≤12, we obtain the upper bounds for Bλ1,λ2m((s−x);x,y) and Bλ1,λ2m((s−x)2;x,y).
Bλ1,λ2m((s−x);x,y)=λ1m(m−1)[1−(1−x)m+1−(x+y)m+1+ym+1−2x(1−(x+y)m+1)]≤1−(1−x)m+1−(x+y)m+1+ym+1+2x(1−(x+y)m+1)2m(m−1):=ϕm(x,y), | (3.23) |
Bλ1,λ2m((s−x)2;x,y)≤x(1−x)m+1m(m−1)[2x2(1−x−y)(x+y)m−1+x(x+y)m+1−xym+1+x(1−x)m−1+x(x+y)m]+1−(1−x)m+1−(x+y)m+1+ym+12m2(m−1):=ψm(x,y). | (3.24) |
Similar steps yield the bounds for Bλ1,λ2m((t−y);x,y) and Bλ1,λ2m((t−y)2;x,y).
Bλ1,λ2m((t−y);x,y)≤1−(1−y)m+1−(x+y)m+1+xm+1+2y(1−(x+y)m+1)2m(m−1)=ϕm(y,x), | (3.25) |
Bλ1,λ2m((t−y)2;x,y)≤y(1−y)m+1m(m−1)[2y2(1−x−y)(x+y)m−1+y(x+y)m+1−yxm+1+y(1−y)m−1+y(x+y)m]+1−(1−y)m+1−(x+y)m+1+xm+12m2(m−1):=ψm(y,x). | (3.26) |
Remark 3.2. By leveraging the endpoint properties of p(λ1,λ2)m,i,j(x,y), we can readily deduce the endpoint interpolation properties of Bλ1,λ2m(f;x,y).
Bλ1,λ2m(f;0,0)=f(0,0),Bλ1,λ2m(f;1,0)=f(1,0),Bλ1,λ2m(f;0,1)=f(0,1). | (3.27) |
We begin by presenting a Korovkin-type approximation result for the bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y).
Theorem 4.1. For any f∈C[S] and |λ1|≤12, |λ2|≤12, the bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y) converge uniformly to f on S.
Proof. Due to the linearity and positivity of Bλ1,λ2m(f;x,y), Lemma 3.1 guarantees that for all (x,y)∈S,
limm→∞Bλ1,λ2m(1;x,y)=1,limm→∞Bλ1,λ2m(s;x,y)=x,limm→∞Bλ1,λ2m(t;x,y)=y,limm→∞Bλ1,λ2m(s2+t2;x,y)=x2+y2, |
all holding uniformly. Applying the Korovkin-Type theorem from [17], it follows that Bλ1,λ2m(f;x,y) converges uniformly to f on S for any f∈C[S].
Next, we analyze the convergence rate of Bλ1,λ2m(f;x,y) approaching to f(x,y). We begin by introducing several smoothness moduli for bivariate functions from [18]. For functions f∈C[S],
● Complete modulus of continuity:
ω(f;α,β)=sup{|f(x1,y1)−f(x2,y2)|:(x1,y1),(x2,y2)∈S,|x1−x2|≤α,|y1−y2|≤β}, |
● Modulus of continuity:
ω(f;δ)=sup{|f(x1,y1)−f(x2,y2)|:(x1,y1),(x2,y2)∈S,√(x1−x2)2+(y1−y2)2≤δ}, |
● Partial moduli of continuity:
ωx(f;α)=sup{sup{|f(x1,y)−f(x2,y)|:|x1−x2|≤α},(x1,y),(x2,y)∈S}, |
ωy(f;β)=sup{sup{|f(x,y1)−f(x,y2)|:|y1−y2|≤β},(x,y1),(x,y2)∈S}, |
where α>0,β>0,δ>0. The following inequalities hold for the aforementioned smoothness moduli (cf. [18]):
ωx(f;γα)≤(1+γ)ωx(f;α), ωy(f;γβ)≤(1+γ)ωy(f;β),γ>0; | (4.1) |
ω(f;γ1α,γ2β)≤(1+γ1+γ2)ω(f;α,β),γ1>0,γ2>0; | (4.2) |
ω(f;α,β)≤ωx(f;α)+ωy(f;β); | (4.3) |
|f(x1,y1)−f(x2,y2)|≤ω(f;|x1−x2|,|y1−y2|)≤ω(f;α,β)≤ω(f;√α2+β2), | (4.4) |
when |x1−x2|≤α,|y1−y2|≤β.
Theorem 4.2. For any f∈C[S] and |λ1|≤12, |λ2|≤12, we have
|Bλ1,λ2m(f;x,y)−f(x,y)|≤3ω(f;√ψm(x,y),√ψm(y,x)). | (4.5) |
Proof. Consider the nonnegativity of p(λ1,λ2)m,i,j(x,y) and Bλ1,λ2m(1;x,y)=1. This implies
|Bλ1,λ2m(f;x,y)−f(x,y)|≤Bλ1,λ2m(|f(s,t)−f(x,y)|;x,y). |
By the inequalities (4.2) and (4.4),
|f(s,t)−f(x,y)|≤ω(f;|s−x|αα,|t−y|ββ)≤(1+|s−x|α+|t−y|β)ω(f;α,β), |
for α, β>0. Thus,
|Bλ1,λ2m(f;x,y)−f(x,y)|≤[1+1αBλ1,λ2m(|s−x|;x,y)+1βBλ1,λ2m(|t−y|;x,y)]ω(f;α,β). |
Applying Cauchy-Schwarz inequalities, we get
Bλ1,λ2m(|s−x|;x,y)≤√Bλ1,λ2m((s−x)2;x,y)≤√ψm(x,y), | (4.6) |
Bλ1,λ2m(|t−y|;x,y)≤√Bλ1,λ2m((t−y)2;x,y)≤√ψm(y,x). | (4.7) |
Hence,
|Bλ1,λ2m(f;x,y)−f(x,y)|≤(1+√ψm(x,y)α+√ψm(y,x)β)ω(f;α,β). |
Setting α=√ψm(x,y),β=√ψm(y,x) yields the desired result
|Bλ1,λ2m(f;x,y)−f(x,y)|≤3ω(f;√ψm(x,y),√ψm(y,x)). |
Theorem 4.3. For any f∈C[S] and |λ1|≤12, |λ2|≤12, we have
|Bλ1,λ2m(f;x,y)−f(x,y)|≤2[ωx(f;√ψm(x,y))+ωy(f;√ψm(y,x))]. | (4.8) |
Proof. To begin, we have
|Bλ1,λ2m(f;x,y)−f(x,y)|≤Bλ1,λ2m(|f(s,t)−f(x,y)|;x,y). |
Utilizing the inequalities (4.1), (4.3), and (4.4), we get
|f(s,t)−f(x,y)|≤ω(f;|s−x|αα,|t−y|ββ)≤ωx(f;|s−x|αα)+ωy(f;|t−y|ββ)≤(1+|s−x|α)ωx(f;α)+(1+|t−y|β)ωy(f;β). |
According to (4.6) and (4.7), we obtain
|Bλ1,λ2m(f;x,y)−f(x,y)|≤[1+Bλ1,λ2m(|s−x|;x,y)α]ωx(f;α)+[1+Bλ1,λ2m(|t−y|;x,y)β]ωy(f;β)≤[1+√ψm(x,y)α]ωx(f;α)+[1+√ψm(y,x)β]ωy(f;β). |
Setting α=√ψm(x,y),β=√ψm(y,x) yields the desired result
|Bλ1,λ2m(f;x,y)−f(x,y)|≤2[ωx(f;√ψm(x,y))+ωy(f;√ψm(y,x))]. |
For f∈C[S], the Peetre's K-functional measures smoothness and is defined as:
K(f;δ)=infg∈C2[S]{‖f−g‖+δ‖g‖C2[S],δ>0}, |
where
● C2[S]:={f∈C[S]:f(i,j)∈C[S],0≤i+j≤2}, f(i,j) is (i,j)th-order partial derivative with respect to x,y of f.
● ‖f‖=sup(x,y)∈S|f(x,y)|.
● ‖f‖C2[S]=‖f‖+∑2i=1(‖∂if∂xi‖+‖∂if∂yi‖).
It is known ([19], page 192) that a constant C>0 exists such that
K(f;δ)≤C[ω2(f;√δ)+min(1,δ)‖f‖], | (4.9) |
for all f∈C[S], where ω2(f;√δ) is the second order modulus of continuity. This property lays the groundwork for a local approximation theorem for Bλ1,λ2m(f;x,y).
Theorem 4.4. For any f∈C[S] and |λ1|≤12, |λ2|≤12, we have
|Bλ1,λ2m(f;x,y)−f(x,y)|≤C[ω2(f;√Θm(x,y)2)+min{1,Θm(x,y)4}‖f‖C2[S]]+ω(f;√ϕ2m(x,y)+ϕ2m(y,x)), | (4.10) |
where Θm(x,y)=ψm(x,y)+ϕ2m(x,y)+ψm(y,x)+ϕ2m(y,x).
Proof. Let
ξλ1m(x,y)=Bλ1,λ2m(s;x,y)=x+λ11−2x−(1−x)m+1−(x+y)m+1+ym+1+2x(x+y)m+1m(m−1),ηλ2m(x,y)=Bλ1,λ2m(t;x,y)=y+λ21−2y−(1−y)m+1−(x+y)m+1+xm+1+2y(x+y)m+1m(m−1). |
Define the auxiliary operators
ˉBλ1,λ2m(f;x,y)=Bλ1,λ2m(f;x,y)−f(ξλ1m(x,y),ηλ2m(x,y))+f(x,y). | (4.11) |
From Lemma 3.1, we have
ˉBλ1,λ2m(1;x,y)=1,ˉBλ1,λ2m(s;x,y)=x,ˉBλ1,λ2m(t;x,y)=y. |
For g∈C2[S] and (s,t)∈S, Taylor's expansion yields
g(s,t)−g(x,y)=∂g(x,y)∂x(s−x)+∫sx(s−u)∂2g(u,y)∂u2du+∂g(x,y)∂y(t−y)+∫ty(t−v)∂2g(x,v)∂v2dv. | (4.12) |
Applying operator ˉBλ1,λ2m to both sides of (4.12), we get
ˉBλ1,λ2m(g;x,y)−g(x,y)=∂g(x,y)∂xˉBλ1,λ2m(s−x;x,y)+ˉBλ1,λ2m(∫sx(s−u)∂2g(u,y)∂u2du;x,y)+∂g(x,y)∂yˉBλ1,λ2m(t−y;x,y)+ˉBλ1,λ2m(∫ty(t−v)∂2g(x,v)∂v2dv;x,y)=ˉBλ1,λ2m(∫sx(s−u)∂2g(u,y)∂u2du;x,y)+ˉBλ1,λ2m(∫ty(t−v)∂2g(x,v)∂v2dv;x,y)=Bλ1,λ2m(∫sx(s−u)∂2g(u,y)∂u2du;x,y)−∫ξλ1m(x,y)x(ξλ1m(x,y)−u))∂2g(u,y)∂u2du+Bλ1,λ2m(∫ty(t−v)∂2g(x,v)∂v2dv;x,y)−∫ηλ2m(x,y)y(ηλ2m(x,y)−v)∂2g(x,v)∂v2dv. |
Applying the triangle inequality and (3.23)–(3.26), we obtain
|ˉBλ1,λ2m(g;x,y)−g(x,y)|≤|Bλ1,λ2m(∫sx(s−u)∂2g(u,y)∂u2du;x,y)|+|∫ξλ1m(x,y)x(ξλ1m(x,y)−u))∂2g(u,y)∂u2du|+|Bλ1,λ2m(∫ty(t−v)∂2g(x,v)∂v2dv;x,y)|+|∫ηλ2m(x,y)y(ηλ2m(x,y)−v)∂2g(x,v)∂v2dv|≤Bλ1,λ2m((s−x)2;x,y)‖g‖C2[S]+(ξλ1m(x,y)−x)2‖g‖C2[S]+Bλ1,λ2m((t−y)2;x,y)‖g‖C2[S]+(ηλ2m(x,y)−y)2‖g‖C2[S]≤[ψm(x,y)+ϕ2m(x,y)+ψm(y,x)+ϕ2m(y,x)]‖g‖C2[S]=Θm(x,y)‖g‖C2[S]. |
On the other hand, we have
|ˉBλ1,λ2m(f;x,y)|≤|Bλ1,λ2m(f;x,y)|+2‖f‖≤‖f‖Bλ1,λ2m(1;x,y)+2‖f‖=3‖f‖. |
Therefore,
|Bλ1,λ2m(f;x,y)−f(x,y)|=|ˉBλ1,λ2m(f;x,y)−f(x,y)+f(ξλ1m(x,y),ηλ2m(x,y))−f(x,y)|≤|ˉBλ1,λ2m(f−g;x,y)|+|f(x,y)−g(x,y)|+|ˉBλ1,λ2m(g;x,y)−g(x,y)|+|f(ξλ1m(x,y),ηλ2m(x,y))−f(x,y)|≤4‖f−g‖+Θm(x,y)‖g‖C2[S]+ω(f;√ϕ2m(x,y)+ϕ2m(y,x)). |
Taking the infimum on the righthand side over all g∈C2[S], we obtain
|Bλ1,λ2m(f;x,y)−f(x,y)|≤4K(f;Θm(x,y)4)+ω(f;√ϕ2m(x,y)+ϕ2m(y,x))≤C[ω2(f;√Θm(x,y)2)+min{1,Θm(x,y)4}‖f‖C2[S]]+ω(f;√ϕ2m(x,y)+ϕ2m(y,x)). |
Therefore, Theorem 4.4 is proved.
Theorem 4.5. For any f∈C1[S], we have
|Bλ1,λ2m(f;x,y)−f(x,y)|≤‖f′x‖[ψm(x,y)]1/2+‖f′y‖[ψm(y,x)]1/2. | (4.13) |
Proof. For any fixed point (x,y)∈S, we have
f(s,t)−f(x,y)=∫sxf′u(u,t)du+∫tyf′v(x,v)dv. | (4.14) |
Applying the operator Bλ1,λ2m to both sides of (4.14), we get
|Bλ1,λ2m(f;x,y)−f(x,y)|≤Bλ1,λ2m(|∫sxf′u(u,t)du|;x,y)+Bλ1,λ2m(|∫tyf′v(x,v)dv|;x,y). |
Since
|∫sxf′u(u,t)du|≤‖f′x‖⋅|s−x|,|∫tyf′v(x,v)dv|≤‖f′y‖⋅|t−y|, |
we obtain
|Bλ1,λ2m(f;x,y)−f(x,y)|≤‖f′x‖Bλ1,λ2m(|s−x|;x,y)+‖f′y‖Bλ1,λ2m(|t−y|;x,y)≤‖f′x‖√ψm(x,y)+‖f′y‖√ψm(y,x). |
Therefore, Theorem 4.5 is proved.
Theorem 4.6. For any f∈C2[S], we have the following Voronovskaja-type asymptotic formula:
limm→∞m[Bλ1,λ2m(f;x,y)−f(x,y)]=x(1−x)2f″xx(x,y)+y(1−y)2f″yy(x,y)−xyf″xy(x,y), | (4.15) |
uniformly in S.
Proof. Let (x,y)∈S be fixed. Utilizing the Taylor formula, we get
f(s,t)=f(x,y)+f′x(x,y)(s−x)+f′y(x,y)(t−y)+12f″xx(x,y)(s−x)2+12f″yy(x,y)(t−y)2+f″xy(x,y)(s−x)(t−y)+r(s,t;x,y)[(s−x)2+(t−y)2], | (4.16) |
where r(s,t;x,y) is the Peano form of the remainder satisfying
lim(s,t)→(x,y)r(s,t;x,y)=0. | (4.17) |
Applying operators Bλ1,λ2m to both sides of (4.16), we obtain
Bλ1,λ2m(f;x,y)−f(x,y)=f′x(x,y)Bλ1,λ2m((s−x);x,y)+f′y(x,y)Bλ1,λ2m((t−y);x,y)+12f″xx(x,y)Bλ1,λ2m((s−x)2;x,y)+12f″yy(x,y)Bλ1,λ2m((t−y)2;x,y)+f″xy(x,y)Bλ1,λ2m((s−x)(t−y);x,y)+Bλ1,λ2m(r(s,t;x,y)[(s−x)2+(t−y)2];x,y). |
By the Cauchy-Schwarz inequality, we get
Bλ1,λ2m(r(s,t;x,y)[(s−x)2+(t−y)2];x,y)≤[Bλ1,λ2m(r2(s,t;x,y);x,y)]12[√Bλ1,λ2m((s−x)4;x,y)+√Bλ1,λ2m((t−y)4;x,y)]. |
Hence, combining (3.21), (3.22), and (4.17), we obtain
limm→∞m[Bλ1,λ2m(r(s,t;x,y)[(s−x)2+(t−y)2];x,y)]=0, | (4.18) |
uniformly in S. By using Corollary 3.1 and (4.18), we obtain
limm→∞m[Bλ1,λ2m(f;x,y)−f(x,y)]=x(1−x)2f″xx(x,y)+y(1−y)2f″yy(x,y)−xyf″xy(x,y), | (4.19) |
uniformly in S. This concludes the proof.
This section numerically illustrates the convergence behavior of the bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y).
Let f(x,y)=x2−√7(1−x−y)2−10xy. Figure 2 displays graphs of f(x,y) (yellow) and its approximations by Bλ1,λ2m(f;x,y) (blue) for varying values of m with λ1=0.5,λ2=−0.5. Table 1 shows the corresponding approximation errors, clearly decreasing with increasing m. This confirms the convergence of the operators.
‖B0.5,−0.5m(f;x,y)−f(x,y)‖∞ | ||||
m=2 | m=5 | m=10 | m=20 | m=50 |
1.3750 | 0.5500 | 0.2750 | 0.1375 | 0.0550 |
We investigate the function f(x,y)=sin(4x)+cos(7y). Figure 3 displays graphs of f(x,y) (yellow) and its approximations by Bλ1,λ2m(f;x,y) (blue) with m=10 and different values of λ1 and λ2. Table 2 shows the approximation errors of Bλ1,λ2m(f;x,y) to f(x,y) for m=10 and various parameter combinations (λ1,λ2). Similarly, Table 3 presents the errors for the function f(x,y)=sin(4(1−x−y))+cos(7xy) under identical conditions.
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
−0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.5 | 0.468932 | 0.465395 | 0.461859 | 0.458323 | 0.454787 |
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.204643 | 0.206170 | 0.207799 | 0.209639 | 0.211569 |
−0.25 | 0.206170 | 0.204761 | 0.206391 | 0.208220 | 0.210128 |
0 | 0.207799 | 0.206391 | 0.204982 | 0.206802 | 0.208734 |
0.25 | 0.209639 | 0.208220 | 0.206802 | 0.205435 | 0.207357 |
0.5 | 0.211569 | 0.210128 | 0.208734 | 0.207357 | 0.205994 |
Table 3 reveals that the bivariate λ-Bernstein operators on triangular domain Bλ1,λ210(f;x,y) outperform the standard bivariate Bernstein operators on triangular domain B0,010(f;x,y) in certain cases. For example, when λ1=λ2=−0.25 or −0.5, the L∞ error norm ‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ is smaller than that of B0,010(f;x,y). This observation highlights the potential of Bλ1,λ2m(f;x,y) to achieve better approximation accuracy in specific scenarios.
This paper presents a novel construction of bivariate Bézier basis functions with shape parameters λ1 and λ2 on a triangular domain. These functions exhibit key properties like nonnegativity, partition of unity and endpoint properties. Building upon them, we introduce the new bivariate λ-Bernstein operators on triangular domain Bλ1,λ2m(f;x,y), encompassing the classical bivariate Bernstein operators on triangular domain as a special case (λ1=λ2=0).
We derive explicit estimates for the moments and central moments of Bλ1,λ2m(f;x,y) through direct calculations. Furthermore, we establish the Korovkin-type approximation theorem for these operators. Additionally, we estimate the rate of convergence of Bλ1,λ2m(f;x,y) toward f(x,y), and rigorously prove both the local approximation theorem and the Voronovskaja-type asymptotic formula for this family of operators.
Numerical examples and diverse graphical representations showcase the convergence behavior of Bλ1,λ2m(f;x,y) under various parameter configurations. Notably, in specific scenarios, these operators achieve superior approximation accuracy compared to the standard bivariate Bernstein operators on triangular domain Bm(f;x,y), highlighting their potential for tailored applications.
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
This work is supported by the Natural Science Foundation of Fujian Province of China (Grant No. 2021J011192) and the Fujian Provincial Office of Education Project (Grant No. JAT200494).
All authors declare no conflicts of interest in this paper.
[1] |
X. M. Zeng, F. Cheng, On the rates of approximation of Bernstein type operators, J. Approx. Theory, 109 (2001), 242–256. https://doi.org/10.1006/jath.2000.3538 doi: 10.1006/jath.2000.3538
![]() |
[2] |
V. Gupta, Some approximation properties of q-Durrmeyer operators, Appl. Math. Comput., 197 (2008), 172–178. https://doi.org/10.1016/j.amc.2007.07.056 doi: 10.1016/j.amc.2007.07.056
![]() |
[3] |
G. Nowak, Approximation properties for generalized q-Bernstein polynomials, J. Math. Anal. Appl., 350 (2009), 50–55. https://doi.org/10.1016/j.jmaa.2008.09.003 doi: 10.1016/j.jmaa.2008.09.003
![]() |
[4] |
P. N. Agrawal, V. Gupta, A. S. Kumar, On q-analogue of Bernstein-Schurer-Stancu operators, Appl. Math. Comput., 219 (2013), 7754–7764. https://doi.org/10.1016/j.amc.2013.01.063 doi: 10.1016/j.amc.2013.01.063
![]() |
[5] | X. Chen, J. Tan, Z. Liu, J. Xie, Approximation of functions by a new family of generalized Bernstein operators, J. J. Math. Anal. Appl., 450 (2017), 244–261. https://doi.org/10.1016/j.jmaa.2016.12.075 |
[6] |
V. Gupta, G. Tachev, A. M. Acu, Modified Kantorovich operators with better approximation properties, Numer. Algor., 81 (2019), 125–149. https://doi.org/10.1007/s11075-018-0538-7 doi: 10.1007/s11075-018-0538-7
![]() |
[7] | D. D. Stancu, A method for obtaining polynomials of Bernstein type of two variables, Am. Math. Month., 70. (1963), 260–264. https://doi.org/10.1080/00029890.1963.11990079 |
[8] |
Q. Cai, B. Lian, G. Zhou, Approximation properties of λ-Bernstein operators, J. Inequal. Appl., 2018 (2018), 61. https://doi.org/10.1186/s13660-018-1653-7 doi: 10.1186/s13660-018-1653-7
![]() |
[9] | Z. Ye, X. M. Zeng, Adjustment algorithms for Bézier curve and surface, In: 2010 5th International Conference on Computer Science & Education, 2010, 1712–1716. https://doi.org/10.1109/ICCSE.2010.5593563 |
[10] |
A. M. Acu, N. Manav, D. F. Sofonea, Approximation properties of λ-Kantorovich operators, J. Inequal. Appl., 2018 (2018), 202. https://doi.org/10.1186/s13660-018-1795-7 doi: 10.1186/s13660-018-1795-7
![]() |
[11] |
Q. Cai, The Bézier variant of Kantorovich type λ-Bernstein operators, J. Inequal. Appl., 2018 (2018), 90. https://doi.org/10.1186/s13660-018-1688-9 doi: 10.1186/s13660-018-1688-9
![]() |
[12] |
Q. Cai, G. Zhou, J. Li, Statistical approximation properties of λ-Bernstein operators based on q-integers, Open Math., 17 (2019), 487–498. https://doi.org/10.1515/math-2019-0039 doi: 10.1515/math-2019-0039
![]() |
[13] |
F. Özger, Weighted statistical approximation properties of univariate and bivariate λ-Kantorovich operators, Filomat, 33 (2019), 3473–3486. https://doi.org/10.2298/FIL1911473O doi: 10.2298/FIL1911473O
![]() |
[14] |
Q. Cai, R. Arslan, On a new construction of generalized q-Bernstein polynomials based on shape parameter λ, Symmetry, 13 (2021), 1919. https://doi.org/10.3390/sym13101919 doi: 10.3390/sym13101919
![]() |
[15] |
A. Kumar, Approximation properties of generalized λ-Bernstein-Kantorovich type operators, Rend. Circ. Mat. Palermo Ser. II, 70 (2021), 505–520. https://doi.org/10.1007/s12215-020-00509-2 doi: 10.1007/s12215-020-00509-2
![]() |
[16] |
R. Pratap, The family of λ-Bernstein-Durrmeyer operators based on certain parameters, Math. Found. Comput., 6 (2023), 546–557. https://doi.org/10.3934/mfc.2022038 doi: 10.3934/mfc.2022038
![]() |
[17] |
C. Badea, I. Badea, H. H. Gonska, A test function theorem and approximation by pseudopolynomials, Bull. Austral. Math. Soc., 34 (1986), 53–64. https://doi.org/10.1017/S0004972700004494 doi: 10.1017/S0004972700004494
![]() |
[18] | G. A. Anastassiou, S. Gal, Approximation theory: Moduli of continuity and global smoothness preservation, Boston: Birkhäuser, 2000. https://doi.org/10.1007/978-1-4612-1360-4 |
[19] | P. L. Butzer, H. Berens, Semi-groups of operators and approximation, New York: Springer, 1967. https://doi.org/10.1007/978-3-642-46066-1 |
‖B0.5,−0.5m(f;x,y)−f(x,y)‖∞ | ||||
m=2 | m=5 | m=10 | m=20 | m=50 |
1.3750 | 0.5500 | 0.2750 | 0.1375 | 0.0550 |
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
−0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.5 | 0.468932 | 0.465395 | 0.461859 | 0.458323 | 0.454787 |
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.204643 | 0.206170 | 0.207799 | 0.209639 | 0.211569 |
−0.25 | 0.206170 | 0.204761 | 0.206391 | 0.208220 | 0.210128 |
0 | 0.207799 | 0.206391 | 0.204982 | 0.206802 | 0.208734 |
0.25 | 0.209639 | 0.208220 | 0.206802 | 0.205435 | 0.207357 |
0.5 | 0.211569 | 0.210128 | 0.208734 | 0.207357 | 0.205994 |
‖B0.5,−0.5m(f;x,y)−f(x,y)‖∞ | ||||
m=2 | m=5 | m=10 | m=20 | m=50 |
1.3750 | 0.5500 | 0.2750 | 0.1375 | 0.0550 |
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
−0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.25 | 0.468442 | 0.464906 | 0.461370 | 0.457834 | 0.454298 |
0.5 | 0.468932 | 0.465395 | 0.461859 | 0.458323 | 0.454787 |
‖Bλ1,λ210(f;x,y)−f(x,y)‖∞ | |||||
λ1 | λ2=−0.5 | λ2=−0.25 | λ2=0 | λ2=0.25 | λ2=0.5 |
−0.5 | 0.204643 | 0.206170 | 0.207799 | 0.209639 | 0.211569 |
−0.25 | 0.206170 | 0.204761 | 0.206391 | 0.208220 | 0.210128 |
0 | 0.207799 | 0.206391 | 0.204982 | 0.206802 | 0.208734 |
0.25 | 0.209639 | 0.208220 | 0.206802 | 0.205435 | 0.207357 |
0.5 | 0.211569 | 0.210128 | 0.208734 | 0.207357 | 0.205994 |