AAA Abstract and Applied Analysis 1687-0409 1085-3375 Hindawi Publishing Corporation 598963 10.1155/2013/598963 598963 Research Article Korovkin Second Theorem via B-Statistical A-Summability 0000-0003-4128-0427 Mursaleen M. 1 Kiliçman A. 2 Başar Feyzi 1 Department of Mathematics Aligarh Muslim University Aligarh 202002 India amu.ac.in 2 Department of Mathematics and Institute for Mathematical Research University Putra Malaysia 43400 Serdang Selangor Malaysia upm.edu.my 2013 14 02 2013 2013 21 09 2012 03 01 2013 2013 Copyright © 2013 M. Mursaleen and A. Kiliçman. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Korovkin type approximation theorems are useful tools to check whether a given sequence (Ln)n1 of positive linear operators on C[0,1] of all continuous functions on the real interval [0,1] is an approximation process. That is, these theorems exhibit a variety of test functions which assure that the approximation property holds on the whole space if it holds for them. Such a property was discovered by Korovkin in 1953 for the functions 1, x, and x2 in the space C[0,1] as well as for the functions 1, cos, and sin in the space of all continuous 2π-periodic functions on the real line. In this paper, we use the notion of B-statistical A-summability to prove the Korovkin second approximation theorem. We also study the rate of B-statistical A-summability of a sequence of positive linear operators defined from C2π() into C2π().

1. Introduction and Preliminaries

Let be the set of all natural numbers, K, and Kn={kn:kK}. Then the natural density of K is defined by (1)δ(K)=limn1n|Kn|=limn(C1χK)n, if the limit exists, where the vertical bars indicate the number of elements in the enclosed set, C1=(C,1) is the Cesàro matrix of order 1, and χK denotes the characteristic sequence of K given by (2)(χK)i={0,if  iK,1,if  iK.

A sequence x=(xk) is said to be statistically convergent to L if for every ε>0, the set Kε:={k:|xk-L|ε} has natural density zero (cf. Fast ); that is, for each ε>0, (3)limn1n|{kn:|xk-L|ε}|=0. In this case, we write L=st-limx.   By the symbol st we denote the set of all statistically convergent sequences. Statistical convergence of double sequences is studied in [2, 3].

A matrix A=(ank)n,k=0 is called regular if it transforms a convergent sequence into a convergent sequence leaving the limit invariant. The well-known necessary and sufficient conditions (Silverman-Toeplitz) for A to be regular are

||A||=supnk|ank|<;

limnank=0, for each k;

limnkank=1.

Freedmann and Sember  generalized the natural density by replacing C1 with an arbitrary nonnegative regular matrix A. A subset K of has A-density if (4)δA(K)=limnkKank exists. Connor  and Kolk  extended the idea of statistical convergence to A-statistical convergence by using the notion of A-density.

A sequence x is said to be A-statistically convergent to L if δA(Kϵ)=0 for every ϵ>0. In this case we write stA-limxk=L. By the symbol stA we denote the set of all A-statistically convergent sequences.

In , Edely and Mursaleen generalized these statistical summability methods by defining the statistical A-summability and studied its relationship with A-statistical convergence.

Let A=(aij) be a nonnegative regular matrix. A sequence x is said to be statistically A-summable to L if for every ϵ>0, δ({in:|yi-L|ϵ})=0; that is, (5)limn1n|{in:|yi-L|ϵ}|=0, where yi=Ai(x). Thus x is statistically A-summable to L if and only if Ax is statistically convergent to L. In this case we write L=(A)st-limx=st-limAx. By (A)st we denote the set of all statistically A-summable sequences. A more general case of statistically A-summability is discussed in .

Quite recently, Edely  defined the concept of B-statistical A-summability for nonnegative regular matrices A and B which generalizes all the variants and generalizations of statistical convergence, for example, lacunary statistical convergence , λ-statistical convergence , A-statistical convergence , statistical A-summability , statistical (C,1)-summability , statistical (H,1)-summability , statistical (N-,p)-summability , and so forth.

Let A=(aij) and B=(bnk) be two nonnegative regular matrices. A sequence x=(xk) of real numbers is said to be B-statisticallyA-summable to L if for every ϵ>0, the set K(ϵ)={i:|yi-L|ϵ} has B-density zero, thus (6)δB(K(ϵ))=limnkK(ϵ)bnk=limn(BχK(ϵ))=limnkbnkχK(ϵ)(k)=0, where yi=Ai(x)=jaijxj. In this case we denote by L=(A)stB-limx=stB-limAx. The set of all B-statistically A-summable sequences will be denoted by (A)stB.

Remark 1.

(1) If A=I (unit matrix), then (A)stB is reduced to the set of B-statistically convergent sequences which can be further reduced to lacunary statistical convergence and λ-statistical convergence for particular choice of the matrix B.

(2) If B=(C,1) matrix, then (A)stB is reduced to the set of statistically A-summable sequences.

(3) If A=B=(C,1) matrix, then (A)stB is reduced to the set of statistically (C,1)-summable sequences.

(4) If B=(C,1) matrix and A=(ajk) are defined by (7)ajk={pkPjif0kj,0otherwise, then (A)stB is reduced to the set of statistically (N-,p)-summable sequences, where p=(pk) is a sequence of nonnegative numbers, such that p0>0 and (8)Pj=k=0jpk(j).

(5) If B=(C,1) matrix and A=(ajk) are defined by (9)ajk={1kljif0kj,0otherwise, where lj=k=0j(1/(k+1)), then (A)stB is reduced to the set of statistically (H,1)-summable sequences.

(6) If a sequence is convergent, then it is B-statistically A-summable, since Ax converges and has B-density zero, but not conversely.

(7) The spaces st,stB,(A)st, and (A)stB are not comparable, even if A=B((C,1)).

(8) If a sequence is A-summable, then it is B-statistically A-summable.

(9) If a sequence is bounded and A-statistically convergent, then it is A-summable and hence statistically A-summable (, see Theorem 2.1) and B-statistically A-summable but not conversely.

Example 2.

(1) Let us define A=(aij), B=(bnk), and x=(xk) by (10)aij={1;ifj=i2,0;otherwise,bnk=(10000········1201200000·····13013013000·····140140140140000··1501501501501500····························),xk={1;ifk  isodd,0;ifk  iseven. Then (11)j=1aijxj={1;ifi  isodd,0;ifi  iseven.

Here xst, x(A)st, xstA, and x(A)stA, but     x is B-statistically A-summable to 1, since δB{i:|yi-1|ϵ}=0. On the other hand we can see that x is B-summable and hence x is B-statistically B-summable, A-statistically B-summable, B-statistically convergent, and statistically B-summable.

Let F() denote the linear space of all real-valued functions defined on . Let C() be the space of all functions f continuous on . We know that C() is a Banach space with norm (12)f:=supx|f(x)|,fC().

We denote by C2π() the space of all 2π-periodic functions fC() which is a Banach space with (13)f2π=supt|f(t)|.

The classical Korovkin first and second theorems statewhatfollows [15, 16]:

Theorem I.

Let (Tn) be a sequence of positive linear operators from C[0,1] into F[0,1]. Then limnTn(f,x)-f(x)=0, for all fC[0,1] if and only if   limnTn(fi,x)-ei(x)=0, for i=0,1,2, where e0(x)=1,e1(x)=x, and e2(x)=x2.

Theorem II.

Let (Tn) be a sequence of positive linear operators from C2π() into F(). Then limnTn(f,x)-f(x)=0,  for all fC2π() if and only if   limnTn(fi,x)-fi(x)=0,  for i=0,1,2, where f0(x)=1,f1(x)=cosx, and f2(x)=sinx.

We write Ln(f;x) for Ln(f(s);x), and we say that L is a positive operator if L(f;x)0 for all f(x)0.

The following result was studied by Duman  which is A-statistical analogue of Theorem II.

Theorem A.

Let A=(ank) be a nonnegative regular matrix, and let (Tk) be a sequence of positive linear operators from C2π() into C2π(). Then for all fC2π()(14)stA-limkTk(f;x)-f(x)2π=0 if and only if (15)stA-limkTk(1;x)-12π=0,stA-limkTk(cost;x)-cosx2π=0,stA-limkTk(sint;x)-sinx2π=0.

Recently, Karakuş and Demirci  proved Theorem II for statistical A-summability.

Theorem B.

Let A=(ank) be a nonnegative regular matrix, and let (Tk) be a sequence of positive linear operators from C2π() into C2π(). Then for all fC2π()(16) st -limnkankTk(f;x)-f(x)2π=0 if and only if (17) st -limnkankTk(1;x)-12π=0, st -limnkankTk(cost;x)-cosx2π=0, st -limnkankTk(sint;x)-sinx2π=0.

Several mathematicians have worked on extending or generalizing the Korovkin's theorems in many ways and to several settings, including function spaces, abstract Banach lattices, Banach algebras, and Banach spaces. This theory is very useful in real analysis, functional analysis, harmonic analysis, measure theory, probability theory, summability theory, and partial differential equations. But the foremost applications are concerned with constructive approximation theory which uses it as a valuable tool. Even today, the development of Korovkin-type approximation theory is far frombeingcomplete. Note that the first and the second theorems of Korovkin are actually equivalent to the algebraic and the trigonometric version, respectively, of the classical Weierstrass approximation theorem . Recently, such type of approximation theorems has been proved by many authors by using the concept of statistical convergence and its variants, for example, . Further Korovkin type approximation theorems for functions of two variables are proved in . In [29, 33] authors have used the concept of almost convergence. In this paper, we prove Korovkin second theorem by applying the notion of B-statistical A-summability. We give here an example to justify that our result is stronger than Theorems II, A, and B. We also study the rate of B-statistical A-summability of a sequence of positive linear operators defined from C2π() into C2π().

2. Main Result

Now, we prove Theorem II for B-statistically A-summability.

Theorem 3.

Let A=(ank) and B=(bnk) be nonnegative regular matrices, and let (Tk) be a sequence of positive linear operators from C2π() into C2π(). Then for all fC2π()(18)stB-limnkankTk(f;x)-f(x)2π=0 if and only if (19)stB-limnkankTk(1;x)-12π=0,stB-limnkankTk(cost;x)-cosx2π=0,stB-limnkankTk(sint;x)-sinx2π=0.

Proof.

Since each of 1, cosx, and sinx belongs to C2π(), conditions (19) follow immediately from (18). Let the conditions (19) hold and fC2π(). Let I  be a closed subinterval of length 2π of   . Fix xI. By the continuity of f at x, it follows that for given ε>0 there is a number δ>0, such that for all t(20)|f(t)-f(x)|<ε, whenever |t-x|<δ. Since f is bounded, it follows that (21)|f(t)-f(x)|2f2π, for all t. For all t(x-δ,2π+x-δ], it is well known that (22)|f(t)-f(x)|<ε+2f2πsin2(δ/2)ψ(t), where ψ(t)=sin2((t-x)/2). Since the function fC2π() is 2π-periodic, the inequality (22) holds for t.

Now, operating Tk(1;x) to this inequality, we obtain (23)|Tk(f;x)-f(x)|(ε+|f(x)|)|Tk(1;x)-1|+ε+f2πsin2(δ/2){|Tk(1;x)-1|+|cosx|+f2πsin2δ2×|Tk(cost;x)-cosx|+f2πsin2δ2+|sinx||Tk(sint;x)-sinx|}ε+(ε+|f(x)|+f2πsin2(δ/2))×{|Tk(1;x)-1|+|Tk(cost;x)-cosx|×+|Tk(sint;x)-sinx|}. Now, taking supxI, we get (24)Tk(f;x)-f(x)ε+K(Tk(1;x)-12π+Tk(cost;x)-cosx2πε+K+Tk(sint;x)-sinx2π), where K:=ε+f2π+(f2π/sin2(δ/2)). Now replace Tk(·,x) by kamkTk(·,x) and then by Bm(·,x) in (24) on both sides. For a given r>0 choose ε>0, such that ε<r. Define the following sets (25)D={mn:Bm(f,x)-f(x)2πr},D1={mn:Bm(f1,x)-f02πr-ε3K},D2={mn:Bm(f2,x)-f12πr-ε3K},D3={mn:Bm(f3,x)-f22πr-ε3K}. Then DD1D2D3, and so δB(D)δB(D1)+δB(D2)+δB(D3). Therefore, using conditions (19) we get (18).

This completes the proof of the theorem.

3. Rate of <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M274"><mml:mrow><mml:mi>B</mml:mi></mml:mrow></mml:math></inline-formula>-Statistical <inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M275"><mml:mrow><mml:mi>A</mml:mi></mml:mrow></mml:math></inline-formula>-Summability

In this section, we study the rate of B-statistical A-summability of a sequence of positive linear operators defined from C2π() into C2π().

Definition 4.

Let A=(aij) and B=(bnk) be two nonnegative regular matrices. Let (αn) be a positive nonincreasing sequence. We say that the sequence x=(xk) is B-statistically A-summable to the number L with the rate o(αn) if for every ε>0, (26)limn1αnkK(ϵ)bnk=0, where K(ϵ)={i:|yi-L|ϵ} and yi=Ai(x)=jaijxj as described above. In this case, we write xk-L=(A)stB-o(αn).

As usual we have the following auxiliary result whose proof is standard.

Lemma 5.

Let (αn) and (βn) be two positive nonincreasing sequences. Let x=(xk) and y=(yk) be two sequences, such that xk-L1=(A)stB-o(αn) and yk-L2=(A)stB-o(βn). Then

c(xk-L1)=(A)stB-o(αn), for any scalar c,

(xk-L1)±(yk-L2)=(A)stB-o(γn),

(xk-L1)(yk-L2)=(A)stB-o(αnβn),

where γn=max{αn,bn}.

Now, we recall the notion of modulus of continuity. The modulus of continuity of fC2π(), denoted by ω(f,δ), is defined by (27)ω(f,δ)=sup|x-y|<δ|f(x)-f(y)|.

It is well known that (28)|f(x)-f(y)|ω(f,δ)(|x-y|δ+1).

Then prove the following result.

Theorem 6.

Let (Tk) be a sequence of positive linear operators from C2π() into C2π(). Suppose that

Tk(1;x)-12π=(A)stB-o(αn),

ω(f,λk)=(A)stB-o(βn), where λk=Tk(φx;x) and  φx(y)=sin2((y-x)/2).

Then for all fC2π(), we have (29)Tk(f;x)-f(x)2π=(A)stB-o(γn),where γn=max{αn,βn}.

Proof.

Let fC2π() and x[-π,π]. Using (28), we have (30)|Tk(f;x)-f(x)|Tk(|f(y)-f(x)|;x)+|f(x)||Tk(1;x)-1|Tk(|x-y|δ+1;x)ω(f,δ)+|f(x)||Tk(1;x)-1|Tk(1+π2δ2sin2(y-x2);x)ω(f,δ)+|f(x)||Tk(1;x)-1|(Tk(1;x)+π2δ2Tk(φx;x))ω(f,δ)+|f(x)||Tk(1;x)-1|. Put δ=λk=Tk(φx;x). Hence we get (31)Tk(f;x)-f(x)2πf2πTk(1;x)-12π+(1+π2)ω(f,λk)+ω(f,λk)Tk(1;x)-12πK{Tk(1;x)-12π+ω(f,λk)K+ω(f,λk)Tk(1;x)-12π}, where K=max{f2π,1+π2}. Hence (32)Tk(f;x)-f(x)2πK{Tk(1;x)-12π+ω(f,λk)+ω(f,λk)K×Tk(1;x)pk-12π}. Now, using Definition 4 and Conditions (i) and (ii), we get the desired result.

This completes the proof of the theorem.

4. Example and Concluding Remark

In the following we construct an example of a sequence of positive linear operators satisfying the conditions of Theorem 3 but does not satisfy the conditions of Theorems II, A, and B.

For any n, denote by Sn(f) the nth partial sum of the Fourier series of f; that is, (33)Sn(f)(x)=12a0(f)+k=1nak(f)coskx+bk(f)sinkx.

For any n, write (34)Fn(f):=1n+1k=0nSk(f).

A standard calculation gives that for every t(35)Fn(f;x)12π-ππf(t)1n+1k=0nsin((2k+1)(x-t)/2)sin((x-t)/2)dt=12π-ππf(t)1n+1k=0nsin2((n+1)(x-t)/2)sin2((x-t)/2)dt=12π-ππf(t)φn(x-t)dt, where (36)φn(x):={sin2((n+1)(x-t)/2)(n+1)sin2((x-t)/2)ifxisnotamultipleof2π,n+1ifxisamultipleof2π. The sequence (φn)n is a positive kernel which is called the Fejér kernel, and the corresponding operators Fn,n1, are called the Fejér convolution operators. We have (37)Fn(1;x)=1,Fn(cost;x)=(nn+1)cosx,Fn(sint;x)=(nn+1)sinx.

Note that the Theorems II, A, and B hold for the sequence (Fn). In fact, we have for every fC2π(), (38)limnFn(f)=f.

Let A=(aij),B=(bnk), and x=(xk) be defined as in Example 2. Let Ln:C2π()C2π() be defined by (39)Ln(f;x)=xnFn(f;x). Then x is not statistically convergent, not A-statistically convergent, and not statistically A-summable, but it is B-statistically A-summable to 1. Since x is B-statistically A-summable to 1, it is easy to see that the operator Ln satisfies the conditions (19), and hence Theorem 3 holds. But on the other hand, Theorems II, A, and B do not hold for our operator defined by (39), since x (and so Ln) is not statistically convergent, not A-statistically convergent, and not statistically A-summable.

Hence our Theorem 3 is stronger than all the above three theorems.

Acknowledgments

This joint work was done when the first author visited University Putra Malaysia as a visiting scientist during August 27–September 25, 2012. The author is very grateful to the administration of UPM for providing him local hospitalities.

Fast H. Sur la convergence statistique Colloquium Mathematicum 1951 2 241 244 MR0048548 ZBL0044.33605 Mursaleen Edely O. H. H. Statistical convergence of double sequences Journal of Mathematical Analysis and Applications 2003 288 1 223 231 10.1016/j.jmaa.2003.08.004 MR2019757 ZBL1032.40001 Mohiuddine S. A. Alotaibi A. Mursaleen M. Statistical convergence of double sequences in locally solid Riesz spaces Abstract and Applied Analysis 2012 2012 9 719729 10. 1155/2012/719729 Freedman A. R. Sember J. J. Densities and summability Pacific Journal of Mathematics 1981 95 2 293 305 MR632187 10.2140/pjm.1981.95.293 ZBL0504.40002 Connor J. On strong matrix summability with respect to a modulus and statistical convergence Canadian Mathematical Bulletin. Bulletin Canadien de Mathématiques 1989 32 2 194 198 10.4153/CMB-1989-029-3 MR1006746 ZBL0693.40007 Kolk E. Matrix summability of statistically convergent sequences Analysis 1993 13 1-2 77 83 MR1245744 ZBL0801.40005 Edely O. H. H. Mursaleen M. On statistical A-summability Mathematical and Computer Modelling 2009 49 3-4 672 680 10.1016/j.mcm.2008.05.053 MR2483671 Mursaleen M. Edely O. H. H. Generalized statistical convergence Information Sciences 2004 162 3-4 287 294 10.1016/j.ins.2003.09.011 MR2076242 ZBL1062.40003 Edely O. H. H. B -statistically A-summability Thai Journal of Mathematics. In press 10.1016/j.mcm.2008.05.053 Fridy J. A. Orhan C. Lacunary statistical convergence Pacific Journal of Mathematics 1993 160 1 43 51 MR1227502 10.2140/pjm.1993.160.43 ZBL0794.60012 Mursaleen λ -statistical convergence Mathematica Slovaca 2000 50 1 111 115 MR1764349 ZBL0953.40002 Móricz F. Tauberian conditions, under which statistical convergence follows from statistical summability (C,1) Journal of Mathematical Analysis and Applications 2002 275 1 277 287 10.1016/S0022-247X(02)00338-4 MR1941785 ZBL1021.40002 Móricz F. Theorems relating to statistical harmonic summability and ordinary convergence of slowly decreasing or oscillating sequences Analysis 2004 24 2 127 145 MR2085952 ZBL1051.40006 Móricz F. Orhan C. Tauberian conditions under which statistical convergence follows from statistical summability by weighted means Studia Scientiarum Mathematicarum Hungarica 2004 41 4 391 403 10.1556/SScMath.41.2004.4.3 MR2102343 ZBL1063.40007 Korovkin P. P. On convergence of linear positive operators in the space of continuous functions Doklady Akademii Nauk 1953 90 961 964 MR0057472 Korovkin P. P. Linear Operators and Approximation Theory 1960 Delhi, India Hindustan MR0150565 Duman O. Statistical approximation for periodic functions Demonstratio Mathematica 2003 36 4 873 878 MR2018707 ZBL1065.41041 Karakuş S. Demirci K. Approximation for periodic functions via statistical A-summability Acta Mathematica Universitatis Comenianae 2012 81 2 159 169 Altomare F. Korovkin-type theorems and approximation by positive linear operators Surveys in Approximation Theory 2010 5 92 164 MR2721174 Demirci K. Dirik F. Approximation for periodic functions via statistical σ-convergence Mathematical Communications 2011 16 1 77 84 MR2835955 Edely O. H. H. Mohiuddine S. A. Noman A. K. Korovkin type approximation theorems obtained through generalized statistical convergence Applied Mathematics Letters 2010 23 11 1382 1387 10.1016/j.aml.2010.07.004 MR2718516 ZBL1206.40003 Mohiuddine S. A. Alotaibi A. Mursaleen M. Statistical summability (C,1) and a Korovkin type approximation theorem Journal of Inequalities and Applications 2012 2012 172 10.1186/1029-242X-2012-172 MR2972647 Mursaleen M. Ahmad R. Korovkin type approximation theorem through statistical lacunary summability Iranian Journal of Science and Technology-Science. In press Mursaleen M. Alotaibi A. Statistical summability and approximation by de la Vallée-Poussin mean Erratum: Applied Mathematics Letters 2012 25 665 10.1016/j.aml.2010.10.014 MR2741037 ZBL1216.40003 Mursaleen M. Alotaibi A. Statistical lacunary summability and a Korovkin type approximation theorem Annali dell'Universitá di Ferrara 2011 57 2 373 381 10.1007/s11565-011-0122-8 MR2855368 Mursaleen M. Karakaya V. Ertürk M. Gürsoy F. Weighted statistical convergence and its application to Korovkin type approximation theorem Applied Mathematics and Computation 2012 218 18 9132 9137 10.1016/j.amc.2012.02.068 MR2923012 Radu C. A -summability and approximation of continuous periodic functions Studia. Universitatis Babeş-Bolyai. Mathematica 2007 52 4 155 161 MR2429346 ZBL1199.41137 Srivastava H. M. Mursaleen M. Khan A. Generalized equi-statistical convergence of positive linear operators and associated approximation theorems Mathematical and Computer Modelling 2012 55 9-10 2040 2051 10.1016/j.mcm.2011.12.011 MR2903563 Anastassiou G. A. Mursaleen M. Mohiuddine S. A. Some approximation theorems for functions of two variables through almost convergence of double sequences Journal of Computational Analysis and Applications 2011 13 1 37 46 MR2760210 ZBL1222.41007 Belen C. Mursaleen M. Yildirim M. Statistical A-summability of double sequences and a Korovkin type approximation theorem Bulletin of the Korean Mathematical Society 2012 49 4 851 861 10.4134/BKMS.2012.49.4.851 Demirci K. Dirik F. Four-dimensional matrix transformation and rate of A-statistical convergence of periodic functions Mathematical and Computer Modelling 2010 52 9-10 1858 1866 10.1016/j.mcm.2010.07.015 MR2719600 Mursaleen M. Alotaibi A. Korovkin type approximation theorem for functions of two variables through statistical A-summability Advances in Difference Equations 2012 2012 65 10.1186/1687-1847-2012-65 MR2948728 Mohiuddine S. A. An application of almost convergence in approximation theorems Applied Mathematics Letters 2011 24 11 1856 1860 10.1016/j.aml.2011.05.006 MR2812226 ZBL1252.41022