Hindawi Publishing CorporationJournal of MathematicsVolume 2013, Article ID 426347, 8 pageshttp://dx.doi.org/10.1155/2013/426347
Research ArticlePointwise Analog of the SteIkin Approximation Theorem
WBodzimierz Aenski
Faculty of Mathematics, Computer Science and Econometrics, University of Zielona Góra,Ulica Szafrana 4a, 65-516 Zielona Góra, Poland
Correspondence should be addressed to Włodzimierz Łenski; [email protected]
Received 11 November 2012; Accepted 20 January 2013
Academic Editor: Roberto Renò
Copyright © 2013 Włodzimierz Łenski.This is an open access article distributed under theCreative CommonsAttribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
We show the pointwise version of the Stečkin theorem on approximation by de la Vallée-Poussin means. The result on normapproximation is also derived.
1. Introduction
Let 𝐿𝑝 (1 ≤ 𝑝 < ∞) [𝐶] be the class of all 2𝜋-periodic real-valued functions integrable in the Lebesgue sense with 𝑝thpower (continuous) over 𝑄 = [−𝜋, 𝜋], and let 𝑋𝑝 = 𝐿𝑝 when1 ≤ 𝑝 < ∞ or 𝑋𝑝 = 𝐶 when 𝑝 = ∞.
Let us define the norms of 𝑓 ∈ 𝑋𝑝 as
𝑓 =
𝑓𝑋𝑝
=𝑓 (⋅)
𝑋𝑝
:=
{{{
{{{
{
{∫𝑄
𝑓 (𝑥)
𝑝𝑑𝑥}
1/𝑝
when 1 ≤ 𝑝 < ∞,
sup𝑥∈𝑄
𝑓 (𝑥) when 𝑝 = ∞,
𝑓𝑥,𝛿
=𝑓
𝑋𝑝,𝑥,𝛿=
𝑓 (⋅)𝑋𝑝 ,𝑥,𝛿
:= sup0
2 Journal of Mathematics
=
{{{{{
{{{{{
{
sup0
Journal of Mathematics 3
𝜎𝑛,𝑚𝑓 (𝑥) − 𝑓 (𝑥) ≤ 𝜋2𝐸∘
𝑛−𝑚(𝑓, 𝑥,
𝜋
2𝑛 − 𝑚 + 1)𝑋𝑝
+ 6𝐹∘
𝑛−𝑚,𝑚(𝑓, 𝑥)
𝑋𝑝
+ ∫
𝜋/(𝑚+1)
𝜋/(2𝑛−𝑚+1)
𝐸∘
𝑛−𝑚(𝑓, 𝑥, 𝑡)
𝑋𝑝
𝑡𝑑𝑡
+ 𝐸∘
𝑛−𝑚(𝑓, 𝑥; 0)
𝑋𝑝,
𝜎𝑛,𝑚𝑓 (𝑥) − 𝑓 (𝑥) ≤ (6 + 𝜋
2) 𝐹𝑛−𝑚,𝑚(𝑓, 𝑥)𝑋𝑝
× [1 + ln 𝑛 + 1𝑚 + 1
]
+ 𝐸𝑛−𝑚(𝑓, 𝑥; 0)𝑋𝑝.
(18)
Now, we can present the main result on pointwiseapproximation.
Theorem 2. If 𝑓 ∈ 𝑋𝑝, then, for any positive integer 𝑚 ≤ 𝑛and all real 𝑥,
𝜎𝑛,𝑚𝑓 (𝑥) − 𝑓 (𝑥)
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1
+ 𝐸2𝑛(𝑓, 𝑥; 0)𝑋𝑝.
(19)
Remark 3. Theorem 2, in case 𝑓 ∈ 𝐶, immediately yields theresult of Stečkin [2]. The result of Stečkin type also holds ifinstead of 𝐶 we consider the spaces 𝑋𝑝 with 1 < 𝑝 < ∞.Then, in the proof, we need the Hardy-Littlewood estimate ofthe maximal function.
At every 𝑋𝑝-point 𝑥 of 𝑓
Ω𝑥𝑓(𝛾)𝑋𝑝= 𝑜𝑥 (1) as 𝛾 → 0+ (20)
and thus fromTheorem 1, we obtain the corollarywhich statesthe result of the Tanović-Miller type [3].
Corollary 4. If 𝑓 ∈ 𝑋𝑝, then, for any positive integer 𝑚 ≤ 𝑛at every 𝑋𝑝-point 𝑥 of 𝑓,
𝜎𝑛,𝑚𝑓 (𝑥) − 𝑓 (𝑥) = 𝑜𝑥 (1) [1 + ln
𝑛 + 1
𝑚 + 1] as 𝑛 → ∞.
(21)
3. Auxiliary Results
In order to prove our theorems, we require some lemmas.
Lemma 5. If 𝑇𝑛 is the trigonometric polynomial of the degreeat most 𝑛 of the best approximation of 𝑓 ∈ 𝑋𝑝 with respect tothe norm ‖ ⋅ ‖𝑋𝑝 , then it is also the trigonometric polynomial ofthe degree at most 𝑛 of the best approximation of 𝑓 ∈ 𝑋𝑝 withrespect to the norm ‖ ⋅ ‖𝑋𝑝 ,𝑥,𝛿 for any 𝛿 ∈ [0, 𝜋].
Proof. From the inequalities𝐸𝑛(𝑓, ⋅, 𝛿)𝑋𝑝
𝑋𝑝≥
𝐸∘
𝑛(𝑓, ⋅, 𝛿)
𝑋𝑝𝑋𝑝
=
𝑓 − 𝑇𝑛,𝛿
∘
𝑋𝑝 ,⋅,𝛿
𝑋𝑝=
𝑓 − 𝑇𝑛,𝛿𝑋𝑝
≥𝑓 − 𝑇𝑛
𝑋𝑝= 𝐸𝑛(𝑓)𝑋𝑝
,
𝐸∘
𝑛(𝑓, ⋅, 𝛿)
𝑋𝑝𝑋𝑝
≤
𝑓 − 𝑇𝑛
∘
𝑋𝑝,⋅,𝛿
𝑋𝑝
=𝑓 − 𝑇𝑛
𝑋𝑝= 𝐸𝑛(𝑓)𝑋𝑝
,
(22)
where 𝑇𝑛,𝛿 and 𝑇𝑛 are the trigonometric polynomials of thedegree at most 𝑛 of the best approximation of 𝑓 ∈ 𝑋𝑝 withrespect to the norms ‖ ⋅ ‖∘
𝑋𝑝 ,𝑥,𝛿and ‖ ⋅ ‖𝑋𝑝 , respectively, we
obtain relation𝑓 − 𝑇𝑛,𝛿
𝑋𝑝=
𝑓 − 𝑇𝑛𝑋𝑝
= 𝐸𝑛(𝑓)𝑋𝑝, (23)
whence 𝑇𝑛,𝛿 = 𝑇𝑛 for any 𝛿 ∈ [0, 𝜋] by uniqueness ofthe trigonometric polynomial of the degree at most 𝑛 of thebest approximation of 𝑓 ∈ 𝑋𝑝 with respect to the norm‖ ⋅ ‖𝑋𝑝 (e.g., see [5] p. 96). We can also observe that for such𝑇𝑛 and any ℎ ∈ [0, 𝛿],
𝑓 − 𝑇𝑛
∘
𝑋𝑝 ,𝑥,ℎ= 𝐸∘
𝑛(𝑓, 𝑥, ℎ)
𝑋𝑝≤ 𝐸𝑛(𝑓, 𝑥, 𝛿)𝑋𝑝
≤𝑓 − 𝑇𝑛
𝑋𝑝,𝑥,𝛿.
(24)
Hence,
𝐸𝑛(𝑓, 𝑥, 𝛿)𝑋𝑝=
𝑓 − 𝑇𝑛𝑋𝑝 ,𝑥,𝛿 (25)
and our proof is complete.
Lemma 6. If 𝑛 ∈ N0 and 𝛿 > 0, then 𝐸𝑛(𝑓, 𝑥; 𝛿)𝑋𝑝 isnonincreasing function of 𝑛 and nondecreasing function of 𝛿.These imply that for 𝑚, 𝑛 ∈ N the function 𝐹𝑛,𝑚(𝑓, 𝑥)𝑋𝑝 isnonincreasing function of 𝑛 and 𝑚 simultaneously.
Proof. The first part of our statement follows from theproperty of the norm ‖ ⋅ ‖𝑥,𝛿 and supremum.The second partis a consequence of the calculation
𝐹𝑛,𝑚+1(𝑓, 𝑥)𝑋𝑝
𝐹𝑛,𝑚(𝑓, 𝑥)𝑋𝑝
=𝑚 + 1
𝑚 + 2(1 +
𝐸𝑛(𝑓, 𝑥; 𝜋/ (𝑚 + 2))𝑋𝑝
∑𝑚
𝑘=0𝐸𝑛(𝑓, 𝑥; 𝜋/ (𝑘 + 1))𝑋𝑝
)
≤𝑚 + 1
𝑚 + 2(1 +
𝐸𝑛(𝑓, 𝑥; 𝜋/ (𝑚 + 1))𝑋𝑝
∑𝑚
𝑘=0𝐸𝑛(𝑓, 𝑥; 𝜋/ (𝑚 + 1))𝑋𝑝
)
=𝑚 + 1
𝑚 + 2(1 +
1
𝑚 + 1) = 1.
(26)
Lemma 7. Let 𝑚, 𝑛, 𝑞 ∈ N0 such that 𝑚 ≤ 𝑛 and 𝑞 ≥ 𝑚 + 1. If𝑓 ∈ 𝑋
𝑝, then
𝜎𝑛+𝑞,𝑚𝑓 (𝑥) − 𝜎𝑛,𝑚𝑓 (𝑥)
≤ 𝐾𝐹𝑛−𝑚,𝑚(𝑓, 𝑥)𝑋𝑝
𝑞−1
∑
𝜈=0
1
𝑚 + 𝜈 + 1.
(27)
4 Journal of Mathematics
Proof. It is clear that
𝜎𝑛,𝑚𝑓 (𝑥) =1
𝑚 + 1
𝑛
∑
𝑘=𝑛−𝑚
1
𝜋∫
𝜋
−𝜋
𝑓 (𝑥 + 𝑡)𝐷𝑘 (𝑡) 𝑑𝑡
=1
𝜋∫
𝜋
−𝜋
𝑓 (𝑥 + 𝑡) 𝑉𝑛,𝑚 (𝑡) 𝑑𝑡,
(28)
where
𝑉𝑛,𝑚 (𝑡) =1
𝑚 + 1
𝑛
∑
𝑘=𝑛−𝑚
𝐷𝑘 (𝑡) ,
𝐷𝑘 (𝑡) =sin ((2𝑘 + 1) 𝑡/2)
2 sin (𝑡/2).
(29)
Hence, by orthogonality of the trigonometric system,
𝜎𝑛+𝑞,𝑚𝑓 (𝑥) − 𝜎𝑛,𝑚𝑓 (𝑥)
=1
𝜋∫
𝜋
−𝜋
[𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)]
× (𝑉𝑛+𝑞,𝑚 (𝑡) − 𝑉𝑛,𝑚 (𝑡)) 𝑑𝑡
=1
𝜋 (𝑚 + 1)
𝑛
∑
𝑘=𝑛−𝑚
∫
𝜋
−𝜋
[𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)]
× (𝐷𝑘+𝑞 (𝑡) − 𝐷𝑘 (𝑡)) 𝑑𝑡
=1
𝜋 (𝑚 + 1)
×
𝑛
∑
𝑘=𝑛−𝑚
∫
𝜋
−𝜋
[𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)]
×sin ((2𝑘 + 2𝑞 +1) 𝑡/2) − sin ((2𝑘 + 1) 𝑡/2)
2 sin (𝑡/2)𝑑𝑡
=1
𝜋 (𝑚 + 1)
𝑛
∑
𝑘=𝑛−𝑚
∫
𝜋
−𝜋
[𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)]
×sin (𝑞𝑡/2) cos ((2𝑘 + 𝑞1) 𝑡/2)
sin (𝑡/2)𝑑𝑡,
(30)
with trigonometric polynomial𝑇𝑛−𝑚 of the degree at most 𝑛−𝑚 of the best approximation of 𝑓.
Using the notations
𝐼1 = [−𝜋
𝑞,𝜋
𝑞] , 𝐼2 = [−
𝜋
𝑚 + 1, −
𝜋
𝑞] ∪ [
𝜋
𝑞,
𝜋
𝑚 + 1] ,
𝐼3 = [−𝜋, −𝜋
𝑚 + 1] ∪ [
𝜋
𝑚 + 1, 𝜋] ,
(31)
we get
∑ =1
𝜋 (𝑚 + 1)
×
𝑛
∑
𝑘=𝑛−𝑚
(∫𝐼1
+∫𝐼2
+∫𝐼3
)
× [𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)]
×sin (𝑞𝑡/2) cos ((2𝑘 + 𝑞 + 1) 𝑡/2)
sin (𝑡/2)𝑑𝑡
= ∑
1
+∑
2
+∑
3
,
∑
1
≤1
𝜋 (𝑚 + 1)
𝑛
∑
𝑘=𝑛−𝑚
∫𝐼1
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡) 𝑞𝑑𝑡
=𝑞
𝜋∫𝐼1
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡) 𝑑𝑡
≤ 2𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑞)
𝑋𝑝
.
(32)
We next evaluate the sums ∑2 and ∑3 using the partialintegrating and Lemma 5. Thus,
∑
2
≤ ∫𝐼2
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)
𝑡𝑑𝑡
= 2[1
2𝑡∫
𝑡
−𝑡
𝑓 (𝑥 + 𝑢) − 𝑇𝑛−𝑚 (𝑥 + 𝑢) 𝑑𝑢]
𝑡=𝜋/(𝑚+1)
𝑡=𝜋/𝑞
+ 2∫
𝜋/(𝑚+1)
𝜋/𝑞
1
𝑡[
1
2𝑡∫
𝑡
−𝑡
𝑓 (𝑥 + 𝑢) − 𝑇𝑛−𝑚 (𝑥 + 𝑢) 𝑑𝑢]𝑑𝑡
≤ 2𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑚 + 1)𝑋𝑝
+ 2∫
𝑡
−𝑡
1
𝑡𝐸𝑛−𝑚(𝑓, 𝑥; 𝑡)𝑋𝑝
𝑑𝑡
≤ 4𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑚 + 1)𝑋𝑝
[1 + ln𝑞
𝑚 + 1]
≤ 4𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑚 + 1)𝑋𝑝
[1 +
𝑞−1
∑
𝜈=0
1
𝑚 + 𝜈 + 1] ,
∑
3
≤1
𝑚 + 1
× ∫𝐼3
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)
𝑡
×
𝑛
∑
𝑘=𝑛−𝑚
cos(𝑘𝑡 +𝑞 + 1
2𝑡)
𝑑𝑡
Journal of Mathematics 5
≤1
𝑚 + 1
× ∫𝐼3
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)
𝑡
×
2 sin((𝑛 + 1) 𝑡/2) cos ((2𝑛 − 𝑚 + 𝑞 + 1) 𝑡/2)2 sin (𝑡/2)
𝑑𝑡
≤𝜋
𝑚 + 1∫𝐼3
𝑓 (𝑥 + 𝑡) − 𝑇𝑛−𝑚 (𝑥 + 𝑡)
𝑡2𝑑𝑡
=𝜋
𝑚 + 1{2 [
1
2𝑡∫
𝑡
−𝑡
𝑓 (𝑥 + 𝑢)
−𝑇𝑛−𝑚 (𝑥 + 𝑢) 𝑑𝑢]
𝑡=𝜋
𝑡=𝜋/(𝑚+1)
+ 4∫
𝜋
𝜋/(𝑚+1)
1
𝑡2[
1
2𝑡∫
𝑡
−𝑡
𝑓 (𝑥 + 𝑢)
− 𝑇𝑛−𝑚 (𝑥 + 𝑢) 𝑑𝑢] 𝑑𝑡}
≤𝜋
𝑚 + 1{2𝐸𝑛−𝑚(𝑓, 𝑥; 𝜋)𝑋𝑝
+ 4∫
𝜋
𝜋/(𝑚+1)
1
𝑡2𝐸𝑛−𝑚(𝑓, 𝑥; 𝑡)𝑋𝑝
𝑑𝑡}
=2𝜋
𝑚 + 1{𝐸𝑛−𝑚(𝑓, 𝑥; 𝜋)𝑋𝑝
+2∫
𝑚+1
1
𝐸𝑛−𝑚(𝑓, 𝑥; 𝜋/𝑢)𝑋𝑝
𝜋2/𝑢2
𝜋𝑑𝑢
𝑢2}
=2𝜋
𝑚 + 1{𝐸𝑛−𝑚(𝑓, 𝑥; 𝜋)𝑋𝑝
+2
𝜋
𝑚−1
∑
𝑘=0
∫
𝑘+2
𝑘+1
𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑢)𝑋𝑝
𝑑𝑢}
=2𝜋
𝑚 + 1{𝐸𝑛−𝑚(𝑓, 𝑥; 𝜋)𝑋𝑝
+2
𝜋
𝑚−1
∑
𝑘=0
𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑘 + 1)𝑋𝑝
}
≤2𝜋 + 2/𝜋
𝑚 + 1
𝑚−1
∑
𝑘=0
𝐸𝑛−𝑚(𝑓, 𝑥;𝜋
𝑘 + 1)𝑋𝑝
,
(33)
which proves Lemma 7.
Before formulating the next lemmas, we define a newdifference. Let 𝑚, 𝑛 ∈ N0 and 𝑚 ≤ 𝑛. Denote that
𝜏𝑛,𝑚𝑓 (𝑥) := (𝑚 + 1) {𝜎𝑛+𝑚+1,𝑚𝑓 (𝑥) − 𝜎𝑛,𝑚𝑓 (𝑥)} . (34)
Lemma 8. Let 𝑚, 𝑛, 𝜇 ∈ N0 such that 2𝜇 ≤ 𝑚 ≤ 𝑛. If 𝑓 ∈ 𝑋𝑝,then
𝜏𝑛,𝑚𝑓 (𝑥) − 𝜏𝑛−𝜇,𝑚−𝜇𝑓 (𝑥)
≤ 𝐾𝜇𝐹𝑛−𝜇+1,𝜇−1(𝑓, 𝑥)𝑋𝑝
ln 𝑚𝜇
.
(35)
Proof. The proof follows by the method of Leindler [6].Namely,
𝜏𝑛,𝑚𝑓 (𝑥) − 𝜏𝑛−𝜇,𝑚−𝜇𝑓 (𝑥)
= (
𝑛+𝑚+1
∑
𝑘=𝑛+𝑚−2𝜇+2
− 2
𝑛
∑
𝑘=𝑛−𝜇+1
)[𝑆𝑘𝑓 (𝑥) − 𝑓 (𝑥)] ,
𝜏𝑛,𝑚𝑓 (𝑥) − 𝜏𝑛−𝜇,𝑚−𝜇𝑓 (𝑥)
≤
(
𝑛+𝑚−𝜇+1
∑
𝑘=𝑛+𝑚−2𝜇+2
−
𝑛
∑
𝑘=𝑛−𝜇+1
)[𝑆𝑘𝑓 (𝑥) − 𝑓 (𝑥)]
+
(
𝑛+𝑚+1
∑
𝑘=𝑛+𝑚−𝜇+2
−
𝑛
∑
𝑘=𝑛−𝜇+1
)[𝑆𝑘𝑓 (𝑥) − 𝑓 (𝑥)]
= 𝜇𝜎𝑛+𝑚−𝜇+1,𝜇−1𝑓 (𝑥) − 𝜎𝑛,𝜇−1𝑓 (𝑥)
+ 𝜇𝜎𝑛+𝑚+1,𝜇−1𝑓 (𝑥) − 𝜎𝑛,𝜇−1𝑓 (𝑥)
.
(36)
By Lemma 6, for 2𝜇 ≤ 𝑚,𝜏𝑛,𝑚𝑓 (𝑥) − 𝜏𝑛−𝜇,𝑚−𝜇𝑓 (𝑥)
≤ 𝐾𝜇𝐹𝑛−𝜇+1,𝜇−1(𝑓, 𝑥)𝑋𝑝[1 + ln
(𝑛 − 𝜇 + 1) + 𝜇 − 1
𝜇]
+ 𝐾𝜇𝐹𝑛−𝜇+1,𝜇−1(𝑓, 𝑥)𝑋𝑝[1 + ln
𝑚 + 𝜇 − 1
𝜇]
≤ 𝐾𝜇𝐹𝑛−𝜇+1,𝜇−1(𝑓, 𝑥)𝑋𝑝[1 + ln 𝑚
𝜇] ,
(37)
and our proof is complete.
Lemma 9. Let 𝑚, 𝑛 ∈ N0 and 𝑚 ≤ 𝑛. If 𝑓 ∈ 𝑋𝑝, then
𝜏𝑛,𝑚𝑓 (𝑥) ≤ 𝐾
𝑛
∑
𝑘=𝑛−𝑚
𝐹𝑘,𝑘−𝑛+𝑚(𝑓, 𝑥)𝑋𝑝. (38)
Proof. Our proof runs parallel with the proof ofTheorem 1 in[2].
If 𝑚 = 0, then𝜏𝑛,0𝑓 (𝑥)
=𝜎𝑛+1,0𝑓 (𝑥) − 𝜎𝑛,0𝑓 (𝑥)
≤ 𝐾𝐹𝑛,0(𝑓, 𝑥)𝑋𝑝, (39)
and if 𝑚 = 1, then𝜏𝑛,1𝑓 (𝑥)
≤ 2𝜎𝑛+2,1𝑓 (𝑥) − 𝜎𝑛,1𝑓 (𝑥)
≤ 𝐾𝐹𝑛−1,1(𝑓, 𝑥)𝑋𝑝
6 Journal of Mathematics
≤ 𝐾 [𝐹𝑛−1,1(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−1,1(𝑓, 𝑥)𝑋𝑝
]
≤ 𝐾 [𝐹𝑛−1,1(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛,1(𝑓, 𝑥)𝑋𝑝
]
(40)
by Lemmas 6 and 7.Next, we construct the same decreasing sequence (𝑚𝑠) of
integers that was given by Stečkin. Let
𝑚0 = 𝑚, 𝑚𝑠 = 𝑚𝑠−1 − [𝑚𝑠−1
2] (𝑠 = 1, 2, . . .) , (41)
where [𝑦] denotes the integral part of 𝑦. It is clear that thereexists smallest index 𝑡 ≥ 1 such that 𝑚𝑡 = 1 and
𝑚 = 𝑚0 > 𝑚1 > ⋅ ⋅ ⋅ > 𝑚𝑡 = 1. (42)
By the definition of the numbers 𝑚𝑠, we have
𝑚𝑠 ≥𝑚𝑠−1
2,
𝑚𝑠−1 − 𝑚𝑠 = [𝑚𝑠−1
2] ≥ [
𝑚𝑠−1
3] (𝑠 = 1, 2, . . . , 𝑡)
(43)
whence
𝑚𝑡−1 = 2, 𝑚𝑡−1 − 𝑚𝑡 = 1,
𝑚𝑠−1 − 𝑚𝑠 ≤ 𝑚𝑠 ≤ 3 (𝑚𝑠 − 𝑚𝑠+1) (𝑠 = 1, 2, . . . , 𝑡 − 1)
(44)
follow.Under these notations, we get the following equality:
𝜏𝑛,𝑚𝑓 (𝑥) =
𝑡
∑
𝑠=1
(𝜏𝑛−𝑚+𝑚𝑠−1,𝑚𝑠−1𝑓 (𝑥) − 𝜏𝑛−𝑚+𝑚𝑠 ,𝑚𝑠
𝑓 (𝑥))
+ 𝜏𝑛−𝑚+𝑚𝑡 ,𝑚𝑡𝑓 (𝑥) ,
(45)
whence, by 𝑚𝑡 = 1,
𝜏𝑛,𝑚𝑓 (𝑥) ≤
𝑡
∑
𝑠=1
𝜏𝑛−𝑚+𝑚𝑠−1,𝑚𝑠−1
𝑓 (𝑥) − 𝜏𝑛−𝑚+𝑚𝑠 ,𝑚𝑠𝑓 (𝑥)
+𝜏𝑛−𝑚+𝑚𝑡 ,𝑚𝑡
𝑓 (𝑥)
(46)
follows.It is easy to see that the terms in the sum ∑𝑡
𝑠=1, by
Lemma 8, with 𝜇 = 𝑚𝑠−1 − 𝑚𝑠 and 𝑚 = 𝑚𝑠−1 do not exceed
𝐾(𝑚𝑠−1 − 𝑚𝑠) 𝐹𝑛−𝑚+𝑚𝑠+1,𝑚𝑠−1−𝑚𝑠(𝑓, 𝑥)
𝑋𝑝ln
𝑚𝑠−1
𝑚𝑠−1 − 𝑚𝑠
,
where (𝑠 = 1, 2, . . . , 𝑡 − 1) ,(47)
and by Lemma 7, we get𝜏𝑛−𝑚+1,1𝑓 (𝑥)
≤ 2𝜎𝑛−𝑚+2,1𝑓 (𝑥) − 𝜎𝑛−𝑚+1,1𝑓 (𝑥)
≤ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝.
(48)
Thus,
𝜏𝑛,𝑚𝑓 (𝑥) ≤ 𝐾
𝑡−1
∑
𝑠=1
3 (𝑚𝑠 − 𝑚𝑠+1) 𝐹𝑛−𝑚+𝑚𝑠+1,𝑚𝑠(𝑓, 𝑥)
𝑋𝑝ln 3
+ 𝐾𝐹𝑛−𝑚+2,𝑚−2(𝑓, 𝑥)𝑋𝑝+ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝
,
(49)
whence, by the monotonicity of 𝐹𝜈,𝜇(𝑓, 𝑥)𝑋𝑝 ,𝜏𝑛,𝑚𝑓 (𝑥)
≤ 𝐾(
𝑡−1
∑
𝑠=1
𝑚𝑠
∑
𝜈=𝑚𝑠+1+1
𝐹𝑛−𝑚+𝜈+1,𝜈(𝑓, 𝑥)𝑋𝑝
+
2
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚−𝜈−1(𝑓, 𝑥)𝑋𝑝)
+ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝
≤ 𝐾
𝑚1+1
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝+ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝
≤ 𝐾
𝑚
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝+ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝
≤ 𝐾
𝑛
∑
𝑘=𝑛−𝑚
𝐹𝑘,𝑘−𝑛+𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐾𝐹𝑛−𝑚,1(𝑓, 𝑥)𝑋𝑝
.
(50)
4. Proofs of the Results
Proof of Theorem 2. The proof follows the lines of the proofsof Theorem 4 in [2] and Theorem in [6]. Therefore, let 𝑛 > 0and 𝑚 ≤ 𝑛 be fixed. Let us define an increasing sequence(𝑛𝑠 : 𝑠 = 0, 1, . . . , 𝑡) of indices introduced by Stečkin inthe following way. Set 𝑛0 = 𝑛. Assuming that the numbers𝑛0, . . . , 𝑛𝑠 are already defined and 𝑛𝑠 < 2𝑛, we define 𝑛𝑠+1 asfollows. Let 𝜈𝑠 denote the smallest natural number such that
𝐹𝑛𝑠−𝑚+𝜈𝑠 ,𝜈(𝑓, 𝑥)
𝑋𝑝≤
1
2𝐹𝑛𝑠−𝑚,𝜈
(𝑓, 𝑥)𝑋𝑝
(𝜈 = 0, 1, . . . , 𝑛) .
(51)
According to the magnitude of 𝜈𝑠, we define
𝑛𝑠+1 =
{{
{{
{
𝑛𝑠 − 𝑚 + 1 for 𝜈𝑠 ≤ 𝑚,𝑛𝑠 + 𝜈𝑠 for 𝑚 + 1 ≤ 𝜈𝑠 < 2𝑛 + 𝑚 − 𝑛𝑠,2𝑛 + 𝑚 for 𝜈𝑠 ≥ 2𝑛 + 𝑚 − 𝑛𝑠.
(52)
If 𝑛𝑠+1 < 2𝑛, we continue the procedure, and if once 𝑛𝑠+1 ≥ 2𝑛,then we stop the construction and define 𝑡 := 𝑠 + 1.
By the previous definition of (𝑛𝑠), we have the followingobvious properties:
𝑡 ≥ 1, 𝑛 = 𝑛0 < 𝑛1 < ⋅ ⋅ ⋅ < 𝑛𝑡, 2𝑛 ≤ 𝑛𝑡 ≤ 2𝑛 + 𝑚,
𝑛𝑠+1 − 𝑛𝑠 ≥ 𝑚 + 1 (𝑠 = 0, 1, . . . , 𝑡 − 1) ,
(53)
Journal of Mathematics 7
and relations
𝐹𝑛𝑠+1−𝑚,𝜈(𝑓, 𝑥)
𝑋𝑝≤
1
2𝐹𝑛𝑠−𝑚,𝜈
(𝑓, 𝑥)𝑋𝑝
for 𝑠 = 0, 1, . . . , 𝑡 − 2,
1
2𝐹𝑛𝑠−𝑚,𝜈
(𝑓, 𝑥)𝑋𝑝
≤ 𝐹𝑛𝑠+1−𝑚−1,𝜈(𝑓, 𝑥)
𝑋𝑝
for 𝑠 = 0, 1, . . . 𝑡 − 1,
(54)
whenever 𝑛𝑠+1 − 𝑛𝑠 > 𝑚 + 1.Let us start with
𝜎𝑛,𝑚𝑓 (𝑥) − 𝑓 (𝑥)
=
𝑡−1
∑
𝑠=0
[𝜎𝑛𝑠 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥)−
𝜎𝑛𝑠+1 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥)]
+𝜎𝑛𝑡 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥)
≤
𝑡−1
∑
𝑠=0
𝜎𝑛𝑠+1,𝑚
𝑓 (𝑥) − 𝜎𝑛𝑠 ,𝑚𝑓 (𝑥)
+
𝜎𝑛𝑡 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥)
=
𝑡−1
∑
𝑠=0
1
𝑚 + 1𝜏𝑛𝑠,𝑚
𝑓 (𝑥)
+
𝜎𝑛𝑡 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥).
(55)
UsingTheorem 1 and that 2𝑛 ≤ 𝑛𝑡 ≤ 2𝑛 + 𝑚, we get
𝜎𝑛𝑡 ,𝑚
𝑓 (𝑥) − 𝑓 (𝑥)
≤ 𝐾𝐹𝑛𝑡−𝑚,𝑚(𝑓, 𝑥)
𝑋𝑝[1 + ln
𝑛𝑡 + 1
𝑚 + 1]
+𝑓 (𝑥) − 𝑇𝑛𝑡−𝑚
(𝑥)
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1+
𝑓 (𝑥) − 𝑇𝑛𝑡−𝑚
(𝑥)
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1
+𝑓 (𝑥) − 𝑇𝑛𝑡−𝑚
(𝑥)
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1
+ 𝐸𝑛𝑡−𝑚(𝑓, 𝑥; 0)
𝑋𝑝
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1
+ 𝐸2𝑛(𝑓, 𝑥; 0)𝑋𝑝.
(56)
The estimate of the sum in the right hand side of (55) wederive from the following one
1
𝑚 + 1𝜏𝑛𝑠,𝑚
𝑓 (𝑥)
≤ 𝐾
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
𝐹𝑛𝑠−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑠−𝑚+𝜈,𝑚
(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1.
(57)
We split the proof of this inequality in two parts. If 𝑛𝑠+1−𝑛𝑠 =𝑚 + 1, then by Lemma 9,
1
𝑚 + 1𝜏𝑛𝑠 ,𝑚
𝑓 (𝑥)
≤ 𝐾
1
𝑚 + 1
𝑛𝑠
∑
𝑘=𝑛𝑠−𝑚
𝐹𝑘,𝑘−𝑛𝑠+𝑚(𝑓, 𝑥)
𝑋𝑝
≤ 𝐾
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
𝐹𝑛𝑠−𝑚+𝜈,𝜈(𝑓, 𝑥)
𝑋𝑝
𝑚 + 𝜈 + 1.
(58)
If 𝑛𝑠+1 − 𝑛𝑠 > 𝑚 + 1, then, by Lemma 7,
1
𝑚 + 1𝜏𝑛𝑠,𝑚
𝑓 (𝑥)
≤ 𝐾𝐹𝑛𝑠−𝑚,𝑚(𝑓, 𝑥)
𝑋𝑝
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
1
𝑚 + 𝜈 + 1,
(59)
and since (1/2)𝐹𝑛𝑠−𝑚,𝑚(𝑓, 𝑥)𝑋𝑝 ≤ 𝐹𝑛𝑠+1−𝑚−1,𝑚(𝑓, 𝑥)𝑋𝑝 , we have
1
𝑚 + 1𝜏𝑛𝑠,𝑚
𝑓 (𝑥)
≤ 2𝐾𝐹𝑛𝑠+1−𝑚−1,𝑚(𝑓, 𝑥)
𝑋𝑝
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
1
𝑚 + 𝜈 + 1
≤ 2𝐾
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
𝐹𝑛𝑠−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝
𝑚 + 𝜈 + 1.
(60)
Consequently,𝑡−1
∑
𝑠=0
1
𝑚 + 1𝜏𝑛𝑠 ,𝑚
𝑓 (𝑥)
≤ 2𝐾
𝑡−1
∑
𝑠=0
𝑛𝑠+1−𝑛𝑠−1
∑
𝜈=0
𝐹𝑛𝑠−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑠−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1.
(61)
Since 𝑛𝑠+1 − 𝑛𝑠 ≤ 2𝑛 + 𝑚 − 𝑛 − 1 = 𝑛 + 𝑚 − 1 for all 𝑠 ≤ 𝑡 − 1,changing the order of summation, we get𝑡−1
∑
𝑠=0
1
𝑚 + 1𝜏𝑛𝑠 ,𝑚
𝑓 (𝑥)
≤ 2𝐾
𝑛+𝑚−1
∑
𝜈=0
1
𝑚 + 𝜈 + 1
× ∑
𝑠:𝑛𝑠+1−𝑛𝑠>𝜈
[𝐹𝑛𝑠−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑠−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
] .
(62)
8 Journal of Mathematics
Using the inequality
𝐹𝑛𝑠+1−𝑚,𝜈(𝑓, 𝑥)
𝑋𝑝≤
1
2𝐹𝑛𝑠−𝑚,𝜈
(𝑓, 𝑥)𝑋𝑝
for { 𝜈 = 0, 1, 2, . . . , 𝑛𝑠+1 − 𝑛𝑠 − 1,𝑠 = 0, 1, 2, . . . , 𝑡 − 2,
(63)
we obtain
∑
𝑠:𝑛𝑠+1−𝑛𝑠>𝜈
[𝐹𝑛𝑠−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑠−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
]
= 𝐹𝑛𝑝−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑝−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
+ ∑
𝑠≥𝑝+1:𝑛𝑠+1−𝑛𝑠>𝜈
[𝐹𝑛𝑠+1−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝
+𝐹𝑛𝑠+1−𝑚+𝜈,𝜈(𝑓, 𝑥)
𝑋𝑝]
≤ 𝐹𝑛𝑝−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑝−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
+ ∑
𝑠:𝑠≥𝑝+1
𝐹𝑛𝑠−𝑚,𝑚(𝑓, 𝑥)
𝑋𝑝[𝐹𝑛𝑠−𝑚,𝑚
(𝑓, 𝑥)𝑋𝑝
+𝐹𝑛𝑠−𝑚,𝜈(𝑓, 𝑥)
𝑋𝑝]
≤ 𝐹𝑛𝑝−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑝−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
+ 2 [𝐹𝑛𝑝+1−𝑚,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑝+1−𝑚,𝜈
(𝑓, 𝑥)𝑋𝑝
]
≤ 3 [𝐹𝑛𝑝−𝑚+𝜈,𝑚(𝑓, 𝑥)
𝑋𝑝+ 𝐹𝑛𝑝−𝑚+𝜈,𝜈
(𝑓, 𝑥)𝑋𝑝
] ,
(64)
where 𝑝 denotes the smallest index 𝑠 having the property𝑛𝑠+1 − 𝑛𝑠 > 𝜈. Hence,
𝑡−1
∑
𝑠=0
1
𝑚 + 1𝜏𝑛𝑠,𝑚
𝑓 (𝑥)
≤ 𝐾
𝑛+𝑚−1
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1
≤ 𝐾
𝑛
∑
𝜈=0
𝐹𝑛−𝑚+𝜈,𝑚(𝑓, 𝑥)𝑋𝑝+ 𝐹𝑛−𝑚+𝜈,𝜈(𝑓, 𝑥)𝑋𝑝
𝑚 + 𝜈 + 1,
(65)
and our proof follows.
References
[1] S. Aljančić, R. Bojanić, andM. Tomić, “On the degree of conver-gence of Fejéer-Lebesgue sums,” L’EnseignementMathématique.Revue Internationale. IIe Série, vol. 15, pp. 21–28, 1969.
[2] S. B. Stečkin, “On the approximation of periodic functions byde la Vallée Poussin sums,” Analysis Mathematica, vol. 4, no. 1,pp. 61–74, 1978.
[3] N. Tanović-Miller, “On some generalizations of the Fejér-Lebesgue theorem,” Unione Matematica Italiana. Bollettino. B.Serie 6, vol. 1, no. 3, pp. 1217–1233, 1982.
[4] W. Łenski, “Pointwise approximation by de la Vallée-Poussinmeans,” East Journal on Approximations, vol. 14, no. 2, pp. 131–136, 2008.
[5] P. L. Butzer and R. J. Nessel, Fourier Analysis and Approxima-tion, Academic Press, New York, NY, USA, 1971.
[6] L. Leindler, “Sharpening of Stečkin’s theorem to strong approx-imation,” Analysis Mathematica, vol. 16, no. 1, pp. 27–38, 1990.
Submit your manuscripts athttp://www.hindawi.com
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttp://www.hindawi.com
Differential EquationsInternational Journal of
Volume 2014
Applied MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Probability and StatisticsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Mathematical PhysicsAdvances in
Complex AnalysisJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
CombinatoricsHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Operations ResearchAdvances in
Journal of
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Decision SciencesAdvances in
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttp://www.hindawi.com
Volume 2014 Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014
Stochastic AnalysisInternational Journal of