Next Article in Journal
Bound Electron Transitions under the Influence of Electromagnetic Wave in Constant Magnetic Field
Previous Article in Journal
A Modified Median String Algorithm for Gene Regulatory Motif Classification
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Definition of Two-Dimensional Schoenberg Type Operators

by
Camelia Liliana Moldovan
and
Radu Păltănea
*,†
Faculty of Mathematics and Computer Science, Transilvania University of Brasov, 500036 Brasov, Romania
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
Symmetry 2020, 12(8), 1364; https://doi.org/10.3390/sym12081364
Submission received: 1 August 2020 / Revised: 12 August 2020 / Accepted: 14 August 2020 / Published: 17 August 2020

Abstract

:
In this paper, a way to build two-dimensional Schoenberg type operators with arbitrary knots or with equidistant knots, respectively, is presented. The order of approximation reached by these operators was studied by obtaining a Voronovskaja type asymptotic theorem and using estimates in terms of second-order moduli of continuity.

1. Introduction

The Schoenberg operators provide a concrete method for obtaining spline approximations of functions. These operators have very good approximation properties. However, they are not very present in the literature. The study of the positive linear Schoenberg operators was the subject of several recent papers, among which we mention here Beutel, Gonska, Kacso and Tachev [1], Tachev [2,3], and Tachev and Zapryanova [4]. In [1] variation-diminishing one-dimensional Schoenberg spline operators, especially with equidistant knots and inequalities in terms of moduli of continuity, were studied. An analysis of the second moment of one-dimensional Schoenberg spline operators moment was presented. A discussion about the degree of simultaneous approximation for multivariate case, more specifically for Boolean sums, was realized. Additionally, a similar discussion was presented for tensor products of one-dimensional Schoenberg spline operators. In [2] a lower bound for the second moment of one-dimensional Schoenberg spline operators is made. In [3] a Voronovskaja’s type theorem for one-dimensional Schoenberg spline operators is presented. In [4] a generalized inverse theorem for one-dimensional Schoenberg spline operators is established.
In practice, the use of the one-dimensional Schoenberg operators offers many advantages. This fact was illustrated, for example, in the recent paper [5], in which these operators were applied for improving the clear sky models used to estimate the direct solar irradiance, with influences of the system design and financial benefits.
The subject of multivariate splines was approached by different methods and from various points of view, such as in the papers written by: Curry and Schoenberg [6], Goodman and Lee [7], de Boor and Hollig [8], Karlin et al. [9], Goodman [10], Chui [11], Schumaker [12], Conti and Morandi [13], Ugarte et al. [14], and Groselj and Knez [15]. Curry and Schoenberg indicated in [6] that the multivariate spline functions can be constructed from volumes of slices of polyhedra; therefore, papers can be found that were written toward that direction. For example, this idea led to the recurrence relations for multivariate splines presented by Karlin et al. in [9]. Goodman and Lee in [7] and Goodman in [10] approached the subject of multidimensional Bernstein–Schoenberg operators depending on m-dimensional volume. In [8] the subject of multidimensional B-splines is treated by de Boor and Hollig as the m-shadow of the polyhedral convex body included in R n . In [11] the Box splines, multivariate truncated powers and many other aspects of multivariate splines are studied by Chui. Conti and Morandi used mixed splines to solve the interproximation problem for surfaces in the case of scattered data in [13]. The aim of Ugarte et al. in the paper [14] was to propose different possibilities of modeling the space–time interaction using one dimensional, two-dimensional and three dimensional B-splines. In [15] Groselj and Knez introduced the notion of a balanced 10-split for the construction of non-negative basis functions for the space of C 1 quadratic splines and showed that the considered split has potential to be used for the construction of C 2 splines.
The aim of our paper is to consider a new approach in spline approximation in the two-dimensional case, based on a two-dimensional version of Schoenberg operators. The two-dimensional Schoenberg type operators are constructed by generalization of the one-dimensional Schoenberg operators. As a result of this generalization we reach a particular form of tensor-product B-splines. The subject of the tensor-product B-splines is treated in several papers, for example in [12].

2. Two-Dimensional Schoenberg Type Operators on Arbitrary Nodes

We define two-dimensional Schoenberg type operators as follows.
Let us consider the knot sequence Δ n , h
0 = α h = α h + 1 = = α 0 < α 1 < α 2 < < α n = α n + 1 = = α n + h = 1 ,
where n > 0 , h > 0 and the knot sequence Δ m , k
0 = β k = β k + 1 = = β 0 < β 1 < β 2 < < β m = β m + 1 = = β m + k = 1 ,
where m > 0 , k > 0 .
The Greville abscissas associated with knot sequence Δ n , h are
ξ i , h : = α i + 1 + α i + 2 + + α i + h h , with h i n 1 ,
and the Greville abscissas associated with knot sequence Δ m , k are
ζ j , k : = β j + 1 + β j + 2 + + β j + k k , with k j m 1 .
The B-splines N i , h ( α ) depending on Δ n , h are defined in the following mode:
N i , h ( α ) = ( α i + h + 1 α i ) [ α i , α i + 1 , , α i + h + 1 ] ( · α ) + h , h i n 1 ,
and the B-splines N j , k ( β ) depending on Δ m , k , by:
N j , k ( β ) = ( β j + k + 1 β j ) [ β j , β j + 1 , , β j + k + 1 ] ( · β ) + k , k j m 1 .
Remark 1.
If α [ α q , α q + 1 ] and β [ β r , β r + 1 ] with 0 q n 1 , 0 r m 1 , then
N i , h ( α ) = 0 , f o r i < q h o r i q + 1 , a n d N i , h ( α ) 0 , f o r q h i q ,
and
N j , k ( β ) = 0 , f o r j < r k o r j r + 1 , a n d N j , k ( α ) 0 , f o r r k j r .
The following relations are well known:
i = h n 1 N i , h ( α ) = 1 , i = h n 1 ξ i , h N i , h ( α ) = α , f o r α [ 0 , 1 ] .
Analogous relations are fulfilled for N j , k .
This notation— Δ 1 = Δ n , h , Δ 2 = Δ m , k , Δ ˜ = Δ 1 × Δ 2 —is used; i.e.,
Δ ˜ : = Δ ˜ n , m h , k = { ( α i , β j ) , h i n + h , k j m + k } .
Definition 1.
Two-dimensional Schoenberg type operator associated with Δ ˜ has the form
S Δ ˜ φ ( α , β ) = i = h n 1 j = k m 1 N i , h ( α ) N j , k ( β ) φ ( ξ i , h , ζ j , k ) ,
where φ : [ 0 , 1 ] × [ 0 , 1 ] R , ( α , β ) [ 0 , 1 ] 2 .
Remark 2.
By taking into account Remark 1 it follows that if α [ α q , α q + 1 ] , with 0 q n 1 and β [ β r , β r + 1 ] with 0 r m 1 , then
S Δ ˜ φ ( α , β ) = i = q h q j = r k r N i , h ( α ) N j , k ( β ) φ ( ξ i , h , ζ j , k ) .
Remark 3.
Two-dimensional Schoenberg type operators are linear and positive like one-dimensional Schoenberg operators. This follows immediately from the linearity and from the positivity of one-dimensional Schoenberg operators. Additionally, obviously, S Δ ˜ is a polynomial of degree at most h in the variable α and with degree at most k in variable β, on each rectangle [ α i 1 , α i ] × [ β j 1 , β j ] , with 0 i n 1 and 0 j m 1 . Moreover S Δ ˜ is a B-spline in each variable.
Two-dimensional Schoenberg type operators admit partial continuous derivatives on [ 0 , 1 ] × [ 0 , 1 ] , since
i + j α i β j ( S Δ ˜ φ ) ( α , β ) = i = h n 1 i α i N i , h ( α ) j = k m 1 j β j N j , k ( β ) φ ( ξ i , h , ζ j , k ) , i , j 0 .
Further, the functions e 0 , π 1 , π 2 C ( [ 0 , 1 ] 2 ) , defined by e 0 ( α , β ) = 1 , π 1 ( α , β ) = α , π 2 ( α , β ) = β , for ( α , β ) [ 0 , 1 ] 2 are considered. The following propositions result directly from Definition 1 and relations (7).
Proposition 1.
For ( α , β ) [ 0 , 1 ] 2 we have
(i)
( S Δ ˜ e 0 ) ( α , β ) = 1 ;
(ii)
( S Δ ˜ π 1 ) ( α , β ) = α ; ( S Δ ˜ π 2 ) ( α , β ) = β ;
(iii)
S Δ ˜ ( π 1 · π 2 ) ( α , β ) = α β .
In the next proposition the notation e 1 ( t ) = t , t [ 0 , 1 ] is used, and e 0 denotes the constant function equal to 1, on the both sets [ 0 , 1 ] and [ 0 , 1 ] 2 .
Proposition 2.
For ( α , β ) [ 0 , 1 ] 2 we have
(i)
( S Δ ˜ ( π 1 α e 0 ) ) ( α , β ) = 0 ( S Δ ˜ ( π 2 β e 0 ) ) ( α , β ) = 0 ;
(ii)
( S Δ ˜ ( π 1 α e 0 ) ( π 2 β e 0 ) ) ( α , β ) = 0 ;
(iii)
( S Δ ˜ ( π 1 α e 0 ) 2 ) ( α , β ) = ( S Δ 1 ( e 1 α e 0 ) 2 ) ( α ) ;
(iv)
( S Δ ˜ ( π 2 β e 0 ) 2 ) ( α , β ) = ( S Δ 2 ( e 1 β e 0 ) 2 ) ( β ) .
Theorem 1.
For the two-dimensional Schoenberg type operators
( S Δ ˜ φ ) ( α , β ) = i = h n 1 j = k m 1 N i , h ( α ) N j , k ( β ) φ ( ξ i , h , ζ j , k )
to converge uniformly on [ 0 , 1 ] × [ 0 , 1 ] to continuous function φ it is sufficient that for any η > 0
( ξ i , h , ζ j , k ) ( α , β ) < η N i , h ( α ) N j , k ( β ) 1 ,
uniformly for 0 α 1 , 0 β 1 when n and m .
Proof. 
We assume that condition (10) is fulfilled. Since φ is continuous on [ 0 , 1 ] × [ 0 , 1 ] , for ε > 0 , η ε > 0 such that for any ( α 1 , β 1 ) [ 0 , 1 ] × [ 0 , 1 ] and ( α 2 , β 2 ) [ 0 , 1 ] × [ 0 , 1 ] with ( α 2 , β 2 ) ( α 1 , β 1 ) < η ε one has | φ ( α 2 , β 2 ) φ ( α 1 , β 1 ) | < ε 2 . Additionally M > 0 s u c h t h a t | φ ( α , β ) | M , ( α , β ) [ 0 , 1 ] 2 .
Let n ε , m ε N such that:
0 < 1 ( ξ i , h , ζ j , k ) ( α , β ) < η ε N i , h ( α ) N j , k ( β ) < ε 4 M , f o r n n ε , m m ε .
Then, for such n and m, we have
| ( S Δ ˜ φ ) ( α , β ) φ ( α , β ) | = | ( ξ i , h , ζ j , k ) ( α , β ) < η ε ( φ ( ξ i , h , ζ j , k ) φ ( α , β ) ) + ( ξ i , h , ζ j , k ) ( α , β ) η ε ( φ ( ξ i , h , ζ j , k ) φ ( α , β ) ) | ε 2 ( ξ i , h , ζ j , k ) ( α , β ) < η ε + 2 M ( ξ i , h , ζ j , k ) ( α , β ) η ε < ε 2 + 2 M ε 4 M = ε .
 □
The norm of the knot sequence Δ ˜ is given by
Δ ˜ : = Δ 1 + Δ 2 ,
where Δ 1 = m a x i ( α i + 1 α i ) and Δ 2 = m a x i ( β i + 1 β i ) .
A quantitative version of the degree of approximation can be given using the first order modulus of continuity, defined as follows:
ω 1 ( φ , ρ ) : = sup { | φ ( u 1 , u 2 ) φ ( v 1 , v 2 ) | , ( u 1 , u 2 ) , ( v 1 , v 2 ) [ 0 , 1 ] 2 , ( u 1 v 1 , u 2 v 2 ) ρ } ,
where φ C ( [ 0 , 1 ] 2 ) , ρ > 0 .
Theorem 2.
For any φ C ( [ 0 , 1 ] 2 ) , operators S Δ ˜ given in (9) satisfy inequality
( S Δ ˜ φ ) φ ω 1 ( φ , Δ ˜ ) ,
where = 1 2 max { h + 1 , k + 1 } .
Proof. 
Let the continuous function φ and let ( α , β ) [ 0 , 1 ] 2 . There exist q { 0 , 1 , , n 1 } and r { 0 , 1 , , m 1 } so that ( α , β ) [ α q , α q + 1 ] × [ β r , β r + 1 ] .
Let q h i q and r k j r . Then
α ξ i , h α q + 1 α i + 1 + + α i + h h α q + 1 α q h + 1 + + α q h 1 h ( h + ( h 1 ) + + 1 ) Δ 1 = h + 1 2 Δ 1 ,
and
α ξ i , h α q α i + 1 + + α i + h h α q α q + 1 + + α q + h h 1 h ( 1 + 2 + + h ) Δ 1 = h + 1 2 Δ 1 .
Therefore | α ξ i , h | h + 1 2 Δ 1 . Similarly | β ζ j , k | k + 1 2 Δ 2 . Then
( α , β ) ( ξ i , h , ζ j , k ) | α ξ i , h | + | β ζ j , k | h + 1 2 Δ 1 + k + 1 2 Δ 2 Δ ˜ .
From Remark 2 it results
| ( S Δ ˜ φ ) ( α , β ) φ ( α , β ) | = i = q h q j = r k r N i , h ( α ) N j , k ( β ) φ ( ξ i , h , ζ j , k ) φ ( α , β ) i = q h q j = r k r N i , h ( α ) N j , k ( β ) | φ ( ξ i , h , ζ j , k ) φ ( α , β ) | i = q h q j = r k r N i , h ( α ) N j , k ( β ) ω 1 ( φ , ( α , β ) ( ξ i , h , ζ j , k ) ) ω 1 ( φ , Δ ˜ ) ω 1 ( φ , Δ ˜ ) .
 □
Corollary 1.
If
Δ ˜ 0 ,
then two-dimensional Schoenberg type operators
( S Δ ˜ φ ) ( α , β ) = i = h n 1 j = k m 1 N i , h ( α ) N j , k ( β ) φ ( ξ i , h , ζ j , k )
converge uniformly on [ 0 , 1 ] × [ 0 , 1 ] to φ, for any continuous function φ.
In [16], the subject of the second moment of variation-diminishing splines is approached. The second moment of the second degree Schoenberg one-dimensional operators was established in [17] and of the third degree Schoenberg one-dimensional operators in [18]. Further on, the form of second moment of two-dimensional Schoenberg type operators with h = k = 3 is presented.
Theorem 3.
The second moment of the two-dimensional Schoenberg type operators S Δ ˜ for h = k = 3 , is
S Δ ˜ ( π 1 α e 0 ) 2 + ( π 2 β e 0 ) 2 ( α , β ) = 1 9 [ α q + 3 α q 1 α q + 2 α q · ( α α q ) 3 α q + 1 α q α q + 2 α q 2 α q + 1 α q 1 · ( α α q + 1 ) 3 α q + 1 α q q 1 i < j q + 2 ( α α i ) ( α α j ) + β r + 3 β r 1 β r + 2 β r · ( β β r ) 3 β r + 1 β r β r + 2 β r 2 β r + 1 β r 1 · ( β β r + 1 ) 3 β r + 1 β r r 1 i < j r + 2 ( β β i ) ( β β j ) ]
where ( α , β ) [ α q , α q + 1 ] × [ β r , β r + 1 ] with 0 q n 1 and 0 r m 1 .
Proof. 
By applying the linearity, Proposition 2(iii) and (iv) follow and the result given in [18]. □

3. Two-Dimensional Schoenberg Type Operators with Equidistant Knots

Now the case h = k = 3 , m = n and equidistant knots is analyzed. More precisely, the equidistant knots are α i = i n , 0 i n , and the extra-knots are α 3 = α 2 = α 1 = 0 and α n + 1 = α n + 2 = α n + 3 = 1 , respectively β j = j n , 0 j n , with extra-knots β 3 = β 2 = β 1 = 0 and β n + 1 = β n + 2 = β n + 3 = 1 .
The Greville abscissas are in this case
ξ i , 3 : = α i + 1 + α i + 2 + α i + 3 3 = α i + 2 , i { 3 , , n 1 } { 2 , n 2 } 1 3 n , i = 2 , 3 n 1 3 n i = n 2
respectively
ζ j , 3 : = β j + 1 + β j + 2 + β j + 3 3 = β j + 2 , j { 3 , , n 1 } { 2 , n 2 } 1 3 n , j = 2 , 3 n 1 3 n j = n 2 .
The B-splines are
N i , 3 ( α ) = ( α i + 4 α i ) [ α i , α i + 1 , α i + 2 , α i + 3 , α i + 4 ] ( · α ) + 3 ,
respectively
N j , 3 ( β ) = ( β j + 4 β j ) [ β j , β j + 1 , β j + 2 , β j + 3 , β j + 4 ] ( · β ) + 3 .
Two-dimensional Schoenberg type operators, with h = k = 3 and m = n , with equidistant knots is denoted by S ˜ n , 3 :
( S ˜ n , 3 φ ) ( α , β ) = i = 3 n 1 j = 3 n 1 N i , 3 ( α ) N j , 3 ( β ) φ ( ξ i , 3 , ζ j , 3 ) ,
and the one-dimensional k degree Schoenberg operators with equidistant knots are denoted by S n , k .
Lemma 1.
The second moment of the two-dimensional Schoenberg type operators S ˜ n , 3 , with n 5 and ( α , β ) [ 0 , 1 ] × [ 0 , 1 ] , verifies the relation
( S ˜ n , 3 ( π 1 α e 0 ) 2 ( α , β ) 1 3 n 2 .
Moreover,
( S ˜ n , 3 ( π 1 α e 0 ) 2 ( α , β ) = 1 3 n 2 ,
for α 2 n , n 2 n and β [ 0 , 1 ] .
Similar relations are true for ( S ˜ n , 3 ( π 2 β e 0 ) 2 ( α , β ) .
Proof. 
The exact form of the second moment of the one-dimensional Schoenberg operators was established in [18]. We have the next cases:
( i ) For χ [ 0 , 1 n ] we have ( S n , 3 ( e 1 χ e 0 ) 2 ) ( χ ) = χ 3 n 18 + χ 3 n with the maximum 2 2 9 n 2 1 3 n 2 ;
( i i ) For χ [ 1 n , 2 n ] we have ( S n , 3 ( e 1 χ e 0 ) 2 ) ( χ ) = χ 3 n 18 χ 2 3 + 2 χ 3 n 1 9 n 2 , which is an increasing function on [ 1 n , 2 n ] with the maximum 1 3 n 2 ;
( i i i ) For χ [ 2 n , n 2 n ] we have ( S n , 3 ( e 1 χ e 0 ) 2 ) ( χ ) = 1 3 n 2 .
By symmetry, the inequality ( S n , 3 ( e 1 χ e 0 ) 2 ) ( χ ) 1 3 n 2 is also obtained for α [ n 2 n , 1 ] .
Finally Proposition 2-(iii) can be applied. In the case of ( S ˜ n , 3 ( π 2 β e 0 ) 2 ( α , β ) Proposition 2-(iv) can be applied. □
Lemma 2.
For n 5 one has
S ˜ n , 3 ( π 1 α e 0 ) 4 + ( π 2 β e 0 ) 4 ( α , β ) 8 3 n 4 .
Proof. 
From [3] we have
( S n , k ( e 1 α e 0 ) 4 ) ( α ) k + 1 2 n 2 ( S n , k ( e 1 α e 0 ) 2 ) ( α ) .
Therefore,
S ˜ n , 3 ( π 1 α e 0 ) 4 + ( π 2 β e 0 ) 4 ( α , β ) 2 n 2 S ˜ n , 3 ( π 1 α e 0 ) 2 + ( π 2 β e 0 ) 2 ( α , β ) 2 · 4 n 2 · 1 3 n 2 = 8 3 n 4 .
 □
The Voronovskaja type theorems are a main topic in studying the convergence properties of the sequences of linear operators. We mention only [3,19,20,21,22]. For two-dimensional Schoenberg type operators S ˜ n , 3 , we obtain the following Voronovskaja type theorem.
Theorem 4.
The following limit is true:
lim n n 2 ( S ˜ n , 3 φ ) ( α , β ) φ ( α , β ) = 1 6 2 φ α 2 ( α , β ) + 2 φ β 2 ( α , β ) .
for any φ C 2 ( [ 0 , 1 ] 2 ) , ( α , β ) ( 0 , 1 ) 2 .
Proof. 
From Taylor’s formula, for any ( γ , θ ) [ 0 , 1 ] × [ 0 , 1 ] it follows that
φ ( θ , γ ) = φ ( α , β ) + φ α ( α , β ) ( θ α ) + φ β ( α , β ) ( γ β ) + 1 2 2 φ α 2 ( α , β ) ( θ α ) 2 + 2 φ α β ( α , β ) ( θ α ) ( γ β ) + 1 2 2 φ β 2 ( α , β ) ( γ β ) 2 + R ( ( θ , γ ) , ( α , β ) ) ,
with the remainder
R ( ( θ , γ ) , ( α , β ) ) = ( θ α , γ β ) 2 g ( θ , γ ) ,
where g ( θ , γ ) 0 when ( θ , γ ) ( α , β ) .
Applying operator S ˜ n , 3 in relation (24) and taking into account Proposition 2 results in
( S ˜ n , 3 φ ) ( θ , γ ) = φ ( α , β ) + 1 2 2 φ α 2 ( α , β ) ( S ˜ n , 3 ( ( π 1 α e 0 ) 2 ) ( α , β ) + 1 2 2 φ β 2 ( α , β ) ( S ˜ n , 3 ( π 2 β e 0 ) 2 ) ( α , β ) + S ˜ n , 3 ( R ( ( θ , γ ) , ( α , β ) ) ) ( α , β ) .
Let ε > 0 . Then there is η > 0 so that | g ( θ , γ ) | < ε if ( θ α , γ β ) < η .
It takes place that
| g ( θ , γ ) | ε + M η 2 ( θ α , γ β ) 2 ,
where ( θ , γ ) [ 0 , 1 ] × [ 0 , 1 ] , when M = g .
We have ( θ α , γ β ) 2 = ( θ α ) 2 + ( γ β ) 2 and ( θ α , γ β ) 4 2 ( θ α ) 4 + 2 ( γ β ) 4 . Then,
| R ( ( θ , γ ) , ( α , β ) ) | ε ( θ α ) 2 + ( γ β ) 2 + 2 M η 2 ( θ α ) 4 + ( γ β ) 4 .
From the relation (28) it follows that
| S ˜ n , 3 R ( ( θ , γ ) , ( α , β ) ) ( α , β ) | S ˜ n , 3 | R ( ( θ , γ ) , ( α , β ) ) | ( α , β ) < ε S ˜ n , 3 ( π 1 α e 0 ) 2 ( α , β ) + S ˜ n , 3 ( π 2 β e 0 ) 2 ( α , β ) + 2 M η 2 S ˜ n , 3 ( π 1 α e 0 ) 4 ( α , β ) + S ˜ n , 3 ( π 2 β e 0 ) 4 ( α , β ) .
From Lemmas 1 and 2 applied in (29), one has
| S ˜ n , 3 R ( ( θ , γ ) , ( α , β ) ) ( α , β ) | ε 2 3 n 2 + 2 M η 2 · 8 3 n 4 .
Therefore,
lim n n 2 S ˜ n , 3 R ( ( θ , γ ) , ( α , β ) ) ( α , β ) = 0 .
Since ( α , β ) ( 0 , 1 ) 2 , for sufficiently great n we have α , β 2 n , n 2 n . Then, for such n from Lemma 1 it follows that ( S ˜ n , 3 ( ( π 1 α e 0 ) 2 ) ( α , β ) = ( S ˜ n , 3 ( ( π 2 β e 0 ) 2 ) ( α , β ) = 1 3 n 2 . Replacing these in (26) and taking into account relation (30), Equation (23) is immediate. □
Moduli of continuity are a powerful tool in evaluating the approximation order. To evaluate the approximation order through the operators S ˜ n , 3 we use general evaluations, expressed with second-order moduli of continuity, demonstrated in [23]. These give a finer evaluation than the evaluations with the first order modulus. For this we introduce the following notation. Let ( X , · X ) be a normed space and D X be a compact and convex set. Let e 0 : X R , e 0 ( t ) = 1 , t D .
If φ C ( X , R ) and h > 0 , then the usual second-order modulus of a function φ C ( X , R ) is defined by
ω ˜ 2 ( φ , h ) : = sup | φ ( u ) 2 f u + v 2 + φ ( v ) | , u , v D , u v X 2 h .
With these elements, a particular version of a more general result given in [23] can be expressed in the form:
Theorem 5.
Let L : C ( X , R ) C ( X , R ) be a positive linear operator. Suppose that X is finite dimensional space with dim X = p . Let x X . Suppose also that
( L ψ ) ( x ) = ψ ( x ) , f o r a l l a f f i n e f u n c t i o n s ψ : X R .
then
| ( L φ ) ( x ) φ ( x ) | p ( L e 0 ) ( x ) + 1 2 h 2 ( L · x X 2 ) ( x ) ω ˜ 2 ( φ , h ) ,
for any φ C ( D , R ) , h > 0 .
In the case of operators S ˜ n , 3 we get:
Theorem 6.
| ( S ˜ n , 3 φ ) ( α , β ) φ ( α , β ) | 2 + 1 3 h 2 n 2 ω ˜ 2 ( φ , h ) ,
where φ C ( [ 0 , 1 ] 2 ) , h > 0 , ( α , β ) [ 0 , 1 ] 2 , n N , n 5 .
Consequently:
( S ˜ n , 3 φ ) φ 7 3 ω ˜ 2 φ , 1 n , φ C ( [ 0 , 1 ] 2 ) , n N , n 5 .
Proof. 
We take X = R 2 , p = 2 , D = [ 0 , 1 ] × [ 0 , 1 ] . Therefore
| ( S ˜ n , 3 φ ) ( α , β ) φ ( α , β ) | 2 ( S n , 3 e 0 ) ( α , β ) + 1 2 h 2 ( S n , 3 ( ( π 1 α e 0 ) 2 + ( π 2 β e 0 ) 2 ) ) ( α , β ) ω ˜ 2 ( φ , h ) = 2 + 1 3 h 2 n 2 ω ˜ 2 ( φ , h ) ,
where h > 0 , ( α , β ) [ 0 , 1 ] × [ 0 , 1 ] .
The particular case is obtained if h = 1 n is chosen. □
In [23] an other second-order global continuity modulus is defined:
ω ˜ 2 ( φ , h ) = s u p { | i = 1 n λ i f ( α i ) φ ( α ) | , α D , α i D , α i α X h , λ i ( 0 , 1 ) , 0 i n , λ 1 + + λ n = 1 } ,
where φ C ( D , R ) , D X is a compact and convex set in the normed space X and h > 0 .
For a more general result given in [23], in a particular case we have the next estimate, which does not depend on the dimension of the space X.
Theorem 7.
Let L : C ( X , R ) C ( X , R ) be a positive linear operator. Suppose that X is finite dimensional space. Let x X . Suppose also that
( L ψ ) ( x ) = ψ ( x ) , f o r a l l a f f i n e f u n c t i o n s ψ : X R ,
then
| ( L φ ) ( x ) φ ( x ) | ( L e 0 ) ( x ) + 1 h 2 ( L · x X 2 ) ( x ) ω ˜ 2 ( φ , h ) ,
for any φ C ( D , R ) , h > 0 .
Applying this theorem to operators S ˜ n , 3 we get:
Theorem 8.
For a function φ continue on [ 0 , 1 ] × [ 0 , 1 ] and h > 0 we have
| ( S ˜ n , 3 φ ) ( α , β ) φ ( α , β ) | 1 + 2 3 n 2 h 2 ω ˜ 2 ( φ , h ) .
Consequently
( S ˜ n , 3 φ ) φ 5 3 ω ˜ 2 φ , 1 n , φ C ( [ 0 , 1 ] × [ 0 , 1 ] ) , n N .
Proof. 
The proof is similar to the proof of Theorem 6. □
Remark 4.
We can make a comparison between the order of approximation reached by the two-dimensional Schoenberg type operators S ˜ n , 3 and that obtained by the two-dimensional Bernstein operators of degree n, which are given by
( B n φ ) ( α , β ) = i = 0 n j = 0 n φ i n , j n p n , i ( α ) p n , j ( β ) , φ : [ 0 , 1 ] 2 R , ( α , β ) [ 0 , 1 ] 2 , n N ,
where p n , i ( t ) = n i t i ( 1 t ) n i , for t = α , β . These operators are the most common approximation polynomial operators.
For these, the following relation is well known:
( B n ( ( π 1 α ) 2 + ( π 2 β ) 2 ) ) ( α , β ) = α ( 1 α ) n + β ( 1 β ) n 1 2 n , ( α , β ) [ 0 , 1 ] 2 , n N .
Thus, if for a function φ : [ 0 , 1 ] 2 R , its values at ( n + 1 ) 2 knots i n , j n , 0 i , j n are known, then applying two-dimensional Bernstein operators we get:
( B n φ ) φ 5 2 ω ˜ 2 φ , 1 2 n , φ C ( [ 0 , 1 ] 2 ) , n N ,
( B n φ ) φ 3 2 ω ˜ 2 φ , 1 2 n , φ C ( [ 0 , 1 ] 2 ) , n N .
On the other hand, if the values of function φ are known at ( n + 3 ) 2 knots ( ξ i , ζ j ) , 3 i , j n 1 , where ξ i , ζ j 0 , 1 3 n , 1 n , 2 n , , n 1 n , 3 n 1 3 n , 1 (see (15) and (16)), then applying two-dimensional Schoenberg operators S ˜ n , 3 given in (19) it results relations (34) and (37). Thus, when using the same tools of measuring for the order of approximations, the advantage is clearly in the favor of Schoenberg operators. Additionally, the volume of computations and the rounding errors are higher in the case of Bernstein operators. Only the smoothness of the image ( B n φ ) is better than the smoothness of ( S ˜ n , 3 φ ) . However, for practical applications, the fact that ( S ˜ n , 3 φ ) has continuous partial derivatives of degree 2 offers a sufficient order of smoothness.

4. Conclusions

In this study, a definition of two-dimensional Schoenberg type operators and their properties has been established. The definition was obtained by generalizing the one-dimensional Schoenberg operators’ formula. The exact forms of the second moment of two-dimensional Schoenberg type operators on arbitrary knots, and on equidistant knots, respectively, alongside a Voronovskaja type theorem, are given. The study presented here also contains estimates with moduli of continuity.
The extension of the Schoenberg operators to the two-dimensional case increases significantly the applicability area of the Schoenberg approximation method. The two-dimensional Schoenberg type operators S ˜ n , 3 generate sufficient smooth surfaces for practical applications and also offer a very good order of approximation of functions. An important advantage of the definition of two-dimensional Schoenberg type operators established in this study consists of the fact that they have a simple form, and this can lead to their easy application in practice.

Author Contributions

All authors equally contributed to this work. All authors have read and agree to this version of the manuscript.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Beutel, L.; Gonska, H.; Kacso, D.; Tachev, G. On variation-diminishing Schoenberg operators: New quantitative statements. Monogr. Acad. Ciencas Zaragoza 2002, 20, 9–58. [Google Scholar]
  2. Tachev, G. A lower bound for the second moment of Schoenberg operator. Gen. Math. 2008, 16, 165–170. [Google Scholar]
  3. Tachev, G. Voronovskaja’s theorem for Schoenberg operator. Math. Ineq. Appl. 2012, 15, 49–59. [Google Scholar] [CrossRef]
  4. Tachev, G.; Zapryanova, T. Generalized Inverse Theorem for Schoenberg Operator. J. Mod. Math. Front. 2012, 1, 11–16. [Google Scholar]
  5. Moldovan, C.L.; Păltănea, R.; Visa, I. Improvement of clear sky models for direct solar irradiance considering turbidity factor variable during the day. Renew. Energy 2020, 161, 559–569. [Google Scholar] [CrossRef]
  6. Curry, H.B.; Schoenberg, I.J. On Polya frequency functions IV: The fundamental spline functions and their limits. J. Anal. Math. 1966, 17, 71–107. [Google Scholar] [CrossRef]
  7. Goodman, T.N.T.; Lee, S.L. Spline Approximation Operators of Bernstein-Schoenberg Type in One and Two Variables. J. Approx. Theory 1981, 33, 248–263. [Google Scholar] [CrossRef] [Green Version]
  8. Boor, C.; Hollig, K. B-Splines from Parallelepipeds; University of Wisconsin-Madison, Mathematics Research Center: Madison, WI, USA, 1982; pp. 1–20. [Google Scholar]
  9. Karlin, S.; Micchelli, C.A.; Rinott, Y. Multivariate Splines: A Probabilistic Perspective. J. Multivariate Anal. 1986, 20, 69–90. [Google Scholar] [CrossRef] [Green Version]
  10. Goodman, T.N.T. Some Properties of Bivariate Bernstein-Schoenberg Operators. Constr. Approx. 1987, 3, 123–130. [Google Scholar] [CrossRef]
  11. Chui, K.C. Multivariate Splines; Society for Industrial and Applied Mathematics: Philadelphia, PA, USA, 1988. [Google Scholar]
  12. Schumaker, L. Spline Functions, 3rd ed.; Cambridge University Press: Cambridge, UK, 2007. [Google Scholar]
  13. Conti, C.; Morandi, R. The bidimensional interproximation problem and mixed splines. J. Comput. Appl. Math. 2001, 130, 1–16. [Google Scholar] [CrossRef] [Green Version]
  14. Ugarte, M.D.; Adin, A.; Goicoa, T. One-dimensional, two-dimensional, and three dimensional B-splines to specify space-time interactions in Bayesian disease mapping: Model fitting and model identifiability. Spat. Stat. 2017, 22, 451–468. [Google Scholar] [CrossRef]
  15. Groselj, J.; Knez, M. A B-spline basis for C1 quadratic splines on triangulations with a 10-split. J. Comput. Appl. Math. 2018, 343, 413–427. [Google Scholar] [CrossRef]
  16. Beutel, L.; Gonska, H.; Kacso, D.; Tachev, G. On the second moments of variation-diminishing splines. J. Concr. Appl. Anal. 2004, 2, 91–117. [Google Scholar]
  17. Moldovan, C.L.; Păltănea, R. Second degree Schoenberg operators with knots at the roots of Chebyshev polynomials. Rev. R. Acad. Cienc. Exactas Fis. Nat. Ser. A Mat. 2019, 113, 2793–2804. [Google Scholar] [CrossRef]
  18. Moldovan, C.L.; Păltănea, R. The Exact Form of the Second Moment of Third Degree Schoenberg Spline Operators. Numer. Funct. Anal. Optim. 2020, 41, 1308–1325. [Google Scholar] [CrossRef]
  19. Gonska, H.; Piţul, P.; Raşa, I. On Peano’s form of the Taylor remainder, Voronovskaja’s theorem and the commutator of positive linear operators. In Numerical Analysis and Approximation Theory, Proceedings of the International Conference 2006, Cluj-Napoca, Romania, 5–8 July 2006; Casa Cartii de Stiinta: Cluj-Napoca, Romania, 2006; pp. 55–80. [Google Scholar]
  20. Gonska, H. On the Degree of approximation in Voronovskaja’s Theorem; Studia Universitatis Babeş-Bolyai Mathematica: Cluj-Napoca, Romania, 2007; Volume LII, pp. 103–115. [Google Scholar]
  21. Gonska, H.; Păltănea, R. General Voronovskaja and Asymptotic Theorems in Simultaneous Approximation. Mediterr. J. Math. 2010, 7, 37–49. [Google Scholar] [CrossRef]
  22. Tachev, G. New estimates in Voronovskaja’s theorem. Numer. Algor. 2012, 59, 119–129. [Google Scholar] [CrossRef]
  23. Păltănea, R. Approximation Theory Using Positive Linear Operators; Birhäuser: Boston, MA, USA, 2004. [Google Scholar]

Share and Cite

MDPI and ACS Style

Moldovan, C.L.; Păltănea, R. A Definition of Two-Dimensional Schoenberg Type Operators. Symmetry 2020, 12, 1364. https://doi.org/10.3390/sym12081364

AMA Style

Moldovan CL, Păltănea R. A Definition of Two-Dimensional Schoenberg Type Operators. Symmetry. 2020; 12(8):1364. https://doi.org/10.3390/sym12081364

Chicago/Turabian Style

Moldovan, Camelia Liliana, and Radu Păltănea. 2020. "A Definition of Two-Dimensional Schoenberg Type Operators" Symmetry 12, no. 8: 1364. https://doi.org/10.3390/sym12081364

APA Style

Moldovan, C. L., & Păltănea, R. (2020). A Definition of Two-Dimensional Schoenberg Type Operators. Symmetry, 12(8), 1364. https://doi.org/10.3390/sym12081364

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop