Date post: | 04-Apr-2018 |
Category: |
Documents |
Upload: | anonymous-7vppkws8o |
View: | 219 times |
Download: | 0 times |
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 1/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
326 | P a g e
Degree Of Approximation Of Functions By Modified Partial Sum
Of Their Conjugate Fourier Series By Generalized Matrix Mean
Prakash Chandra Rautaray1, Ellipse
2
1
Department of Mathematics, KIIT University, Bhubaneswar-751024, Odisha, India 2Department of Mathematics, Maharishi College of Natural Law, Bhubaneswar -751007, Odisha, India
AbstractThe paper studies the degree of
approximation of conjugate of a 2 -periodic
Lebesgue integrable function f by using modified
partial sum of its conjugate Fourier series bygeneralized matrix mean in generalized Holder metric.
Keyword. Banach Space, generalized Holder metric and regular generalized matrix.
1. Definition and NotationThe following definitions will be used
throughout the paper (see Zygmund [8] p.16,42, [5]
p.2,49 and [3] ).
(i) The space , p L includes the space
of all 2 -periodic Lebesgue integrable continuous
functions defined in , with p-norm given by
12
0
2
0
sup ;
; 1
;0 1 .
t
p p
p
p
f t p
f f t dt p
f t dt p
(ii) 0
, supc
h
w w f f x h f x
when p , is called the modulus of continuity
0
, sup p p p
h
w w f f x h f x
is called the integral modulus of continuity.
2 2
0
,
sup 2
p p
ph
w w f
f x h f x h f x
is called the integral modulus of smoothness.(iii) The Lipschitz condition is given by
or
0
supp
h
f x h f x K
(+ve constant )
when 0 . p
(iv) The Holder metric space H is defined by
2 : ; 0,0 1 H f C f x f y K x y K
with Holder metric induced by the norm
,
sup , sup supc
x y t x y
f x f y f f f x y f t
x y
where
, f x f y
f x y x y
and 0 for
0 .
(v) A normed linear space which is completein the metric defined by its norm is called BanachSpace.
(vi) The generalized Holder metric space
, H p is defined by
, : p p H p f L f x h f x K h
where K > 0(constant), 0< 1 and 0 . p
Also the metric given by
,
sup , sup p
p p ph h
f x h f x f f f x h x f
h
and ,o p p
f f for 0 , is called generalized
Holder metric.
(vii) , H p is a complete normed linear space
and hence a Banach space for 0 1. p
Also , . H H
0
sup c
h
f x h f x K
(+ve constant) when p
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 2/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
327 | P a g e
(viii)A generalized matrix nk M m i is said to be regular if
, 0
sup nk n i k
M m i
and 0
1nk
k
m i
as n uniformly in i.
(ix) Let , ; 1, p f L p be a 2 -periodic Lebesgue integrable function.
Then the conjugate Fourier series of f at ,t x is given by
1
sin cos ;k k
k
f x a kx b kx
(1)
where 1
cosk a f t kt dt
and 1
sin .k b f t kt dt
The following notations shall be used throughout the paper.
*
1 cos
2tan2
n
nt D t
t
is called Dirichlet’s modified conjugate kernel.
1
2 x t f x t f x t
, , 1
0
n n k n k
k
i m i m i
0
1 cos;
2tan2
n nk
k
kt K i t m i
t
(2)
*
0
cos;2tan
2
n nk
k
kt H i t m it
(3)
0
2
2tan2
xt
f x dt t
is called conjugate function of f x .
2
, ,2tan
2n
x
n
t f x f x dt
t
; (4)
where 0n
is very small positive number.
* *
0
2n x n
S x t D t dt
(5)
is called modified nth partial sum of conjugate Fourier series of f x given by (1).
* *
0
n nk k
k
M S x m i S x
(6)
uniformly in i, provided the series exists for each n, which is called the matrix
transformation of *
nS x .
* *; ,n n
n
l i x M S x f x
(7)
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 3/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
328 | P a g e
2. IntroductionChandra [1] and Sahney have determined
the degree of approximation of a function
belonging to Lip by ,1 , ,C C and
, n N p means. In 1981, Quereshi discussed the
degree of approximation of conjugate of a function
belonging to Lip and Lip , p by , n N p
means of con jugate series of a Fourier series. In2000, Shyam Lal [4] determined the degree of
approximation of conjugate of function belonging
to weighted class , pW L t by matrix means
of conjugate series of Fourier series. Also in 2001,G.Das, R.N.Das and B.K.Ray[3] studied the degreeof approximation in same direction using infinitematrix mean in generalized Holder metric.
The objective of the present paper is tostudy more comprehensively the result of G.Das,R.N.Das & B.K.Ray[3] by generalized
matrix mean.
3. Result In this paper we have studied the degree of
approximation of conjugate function of f x by
modified partial sum of its conjugate Fourier series
by generalized matrix mean in generalized Holder metric i.e.
* *
,0
,
; ,n nk k p
k n p
l i x m i S x f x
uniformly in i.
The following lemma will be required for establishing the theorem.
Lemma.
Let 0 p .
Then (a) , x p pt w f
and (b) x y x pt t K
p f x t y f x t ;where K > 0 (constant).
Proof .
(a) For 1 p and by Minkowski’s inequality, we have
1 12 2 2
0 0 0
p p p p p f x t f x t dx f x t f x dx f x f x t dx
and for 0 1 p , we have by modified Minkowski’s inequality
2 2 2
0 0 0
p p p f x t f x t dx f x t f x dx f x f x t dx
0
sup p
t
f x t f x t
0 0
sup sup p p
t t
f x t f x f x f x t
0
2 sup p
t
f x t f x
0 0
sup sup2 p
t t p
f x t f x t f x t f x
, x p pt w f
(b) Now
x y xt t = 1
2 f x y t f x y t f x t f x t
= 1 1
2 2 f x t y f x t f x t f x t y
By Minkowski’s inequality for 1 p and 0 1 p separately, we get
1 1
2 2 x y x p p pt t f x t y f x t f x t f x t y
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 4/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
329 | P a g e
p
K f x t y f x t ;where K > 0 is a constant
which establishes the lemma.
Theorem.
If for 1,0 1 p and for positive increasing sequences n and n such that n n and
nk M m i (space of regular matrices) such that 2
1n
nk n
k
k m i O
and , f H p , then
*
1,
1
1log
; 11 ;0 1
log log ; 1
nn n
n n
n p
n n
n
n n n
l i x O
i
uniformly in i.
Proof.
The equation (7) can be written as
* *; ,n n
n
l i x M S x f x
= *
0
,nk k
k n
m i S x f x
=
0 00
1 cos2 2
2 tan 2 tan2 2n
x
nk x nk
k k
kt t m i t dt m i dt
t t
( 0
1nk
k
m i
as n )
=
0 00
1 cos2 2 1
2 tan 2 tan
2 2n
x nk x nk
k k
kt t m i dt t m i dt
t t
(provided the change of order of summation is permitted)
=
0 00
1 cos cos2 2
2 tan 2 tan2 2
n
n
x nk x nk
k k
kt kt t m i dt t m i dt
t t
*
0
2 2; ;
n
n
x n x nt K i t dt t H i t dt
(8)
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 5/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
330 | P a g e
Now
0
1 cos;
2tan2
n nk
k
kt K i t m i
t
0
2 1 cos 2
2tan2
nk
k
m i kt t
0
22 tan 0
2nk
k
t m i t for t
t
2 M
t
1O t
1;n K i t O t (9)
Also for 1,0 1 p and , , f H p
p f x t f x O t
(10)
By lemma(a), 0
x p pt
t Sup f x t f x O t
( by (10))
x pt O t
(11)
Again x y x pt t
x y x p pt t
(by Minkowski’s inequality )
O t O t
( by (11))
x y x p
t t O t
(12)
Consider * *; ;n n
pl i x y l i x
= *
0
2 2; ;
n
n
x y x n x y x n
p
t t K i t dt t t H i t dt
*
0
2 2
; ;
n
n
x y x n x y x n p pt t K i t dt t t H i t dt
= 1 2 I I (say) (13)
where 1
0
2;
n
x y x n p I t t K i t dt
= 1
0
2 n
O t O t dt
(by (12) and (9))
= 1
0
1n
O t dt
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 6/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
331 | P a g e
= 0
1nt
O
= 1
1n
O
(14)
Also by Abel’s transformation for n
t
,
1
, , 1 ,
0 0
0
cos lim ;
1sin cos
2 2
sin2
n
nk k n k n k n n nn
k k
n
n
k
m i kt S m i m i S m i
n nt t
where S cos kt t
(see[2],[8])
, , 1
0
1sin cos2 2
sin2
n k n k
k
k kt t
m i m it
as ,lim 0n nn
m i
, , 1
0 0
1cos
sin2
nk n k n k
k k
m i kt m i m it
1sin cos
12 2
sin sin2 2
k kt t
t t
= 1
, , 1
0
n k n k
k
O t m i m i
= 1
nO t i
1
0
cosnk n
k
m i kt O t i
for
n
t
(15)
and *
2
2;
n
x y x n p I t t H i t dt
=
0
cos2
2tan2n
nk
k
kt O t m i dt
t
( by (12))
= 1
0
2cos
n
nk
k
O t O t m i kt dt
2 tan 0
2
t t for t
= 1 12
n
nO t O t O t i dt
(by (15)
= 21
n
nO i t dt
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 7/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
332 | P a g e
=
1
;0 11 1
log ; 1n
n
t
O i
t
=
1;0 1
1log ; 1
nn
n
O i
(16)
Further order estimates for I1 and I2 can be obtained as follows
1
0
2 2; ;
n n
n
x y x n x y x n p p I t t K i t dt t t K i t dt
= 11 12 I I (say) (17)
Also
0
1 cos;
2tan 2
n nk
k
kt K i t m i
t
2 2
0 1
sin sin2 2
tan tan2 2
n
n
nk nk
k k
kt kt
m i m it t
2 2 2 2
0 14 tan 4 tan2 2
n
n
nk nk
k k
k t k t m i m i
t t
sin2 2
kt kt
2 2
0 12 2
n
n
nk n nk
k k
t t m i m i k
2tan2
t t
= 2 2
0 1
n
n
nk n nk
k k
O t m i k m i
= n nO t O O ( 0
nk
k
m i
and by necessary condition of theorem)
;n n K i t O t O (18)
Again by lemma (b), x y x pt t O y
(19)
Now 11
0
2;
n
x y x n p I t t K i t dt
= 0
2 n
nO y O O t dt
( by (18),(19) )
= 0
n
nO y O tdt
=
2
02
n
n
t O y O
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 8/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
333 | P a g e
= 2
1n
n
O y O O
= 1
n
O y
(20)
and 12
2;
n
n
x y x n p I t t K i t dt
=
0
1 cos2
2tan2
n
n
nk
k
kt O y m i dt
t
( by (19))
=
1 2
0
2sin 2 tan
2 2
n
n
nk
k
kt t O y O t m i dt t
= 1 2
0
2 sin 12
n
n
nk
k
kt O y O t m i dt
= 1n
n
O y M O t dt
= 1
n
n
O y dt t
= log n
n
O y t
log n
n
O y
(21)
Hence 1 11 12
1log n
n n
I I I O y
(22)
Combining (14) and (22), we get for 0 1
1
1 1 1 I I I
= 1
1 1log 1n
n n n
O y O
1log n
n
n n
O y
(23)
Also 2
2,
n
x y x n p I t t K i t dt
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 9/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
334 | P a g e
= 1
0
2
n
nk
k
O y O t m i cos kt dt
( by (19) and 2tan
2
t t )
= 1 1
n
nO y O t O t i dt
( by (15) )
= 2
n
nO y i t dt
= 2
n
nO y i t dt
= 1
1n
n
t O y i
= n n i O y (24)
Combining (16), (24), we have for 0 1
1
2 2 2 I I I
=
111 ;0 1
1 log ; 1
n n
n n
n n
O ii O y
O i
=
1
1
;0 1
log ; 1
n
n
n n
O y i
(25)
Hence * *1 2
; ;n n
pl i x y l i x I I
=
1
1
;0 11log
log ; 1
nnn n
n n n n
O y O y i
* *
1, 0
1
1log
; ;sup 1
;0 1
log ; 1
nn
n n p n n
n yn
n
n n
l i x y l i xO
yi
(26)
Further * ;n
pl i x
*
0
2 2; ;
n
n
x n x n p pt K i t dt t H i t dt
= 1
00
cos2 2
2tan2
n
n
nk
k
kt O t O t dt O t m i dt
t
( using (11) in both integrals & (9) in 1st
integral )
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 10/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
335 | P a g e
= 1 1
00
1 1 cos 2 tan2
n
n
nk
k
t O t dt O t O t m i kt dt t
= 1 1
0
1 1n
n
nt O O t O t i dt
( by (15) )
= 211 1
n
n
n
O O i t dt
=
1
;0 11 1 1
log ; 1n
n n
t
O O i
t
=
1 ;0 11 1
log ; 1nn n
n
O O i
1
* ;0 1; 1
log ; 1
n
n n n p
n
l i x O i
(27)
Adding (26) and (27), we have
* *
* *
, , 0
; ;; ; sup
n n p
n n p p n y
l i x y l i xl i x l i x
y
1
1
1 ;0 11
1 loglog log ; 1
n nn
n n n
n nn n n
O i
(28)
Hence the result follows.
This completes the proof of theorem.
4. CorollariesUsing the above theorem the following two corollaries can be established.
Corollary1
If for 1,0 1, , p f H p and nk M m i is a lower triangular matrix of non-negative real
numbers with monotonic increasing in k such that 0
1n
nk
k
m i
as n uniformly in i, then
1
*
, 1,
1 ;0 1; 1 1
log log ; 1n n n
p
n nl i x O n n m i
n n n
Proof.
Let 1,0 1, , . p f H p
Let nk M m i be a lower triangular matrix of non-negative real numbers with monotonic increasing in k
such that 0
1n
nk
k
m i
as n uniformly in i i.e.
, , 10, 0,1, 2..., 1nk n k n k m i m i m i k n and 0
1
n
nk
k m i
as n uniformly in i.
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 11/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
336 | P a g e
Choose .n n n
Then 2 2
1 1n
nk nk
k k n
k m i k m i
= 2 2
1 0 2 0 ...n n
= O n (29)
Also , , 1 , 1 ,
0 0
n
n n k n k n k n k
k k
i m i m i m i m i
= 1
, 1 , , 1 ,
0
n
n k n k n n n n
k
m i m i m i m i
= 1
, 1 , , , 1
0
0n
n k n k n n n n
k
m i m i m i m i
= ,1 ,0 ,2 ,1 , , 1 ,...n n n n n n n n n nm i m i m i m i m i m i m i
= ,0 , ,n n n n nm i m i m i
= , ,02
n n nm i m i
,2 n nm i
,n n ni O m i (30)
Clearly all conditions of theorem are hold.
Hence
1
*
, 1,
1 ;0 1; 1
log log ; 1n n n
p
n nl i x O n n n O m i
n n n
1
, 1
1 ;0 11 1
log log ; 1n n
n nO n n m i
n n n
(31)
This establishes corollary 1.
Corollary 2
If for 1,0 1, , p f H p and nk M m i is a lower triangular matrix of non-negative real
numbers with monotonic decreasing in k such that 0
1n
nk k
m i
as n uniformly in i, then
1
*
,0 1,
1 ;0 1; 1 1
log log ; 1n n
p
n nl i x O n n m i
n n n
Proof.
Let 1,0 1, , . p f H p
Let nk M m i be a lower triangular matrix of non-negative real numbers with monotonic decreasing in k
such that 0
1n
nk k
m i
as n uniformly in i i.e.
, , 10, 0,1, 2..., 1nk n k n k m i m i m i k n and 0
1n
nk k
m i
as n uniformly in i.
Choose .n n n
Then 2 2
1 1n
nk nk
k k n
k m i k m i
= 2 2
1 0 2 0 ...n n
7/29/2019 Aw 31326337
http://slidepdf.com/reader/full/aw-31326337 12/12
Prakash Chandra Rautaray, Ellipse / International Journal of Engineering Research and
Applications (IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 1, January -February 2013, pp.326-337
337 | P a g e
= O n (32)
Also , , 1
0
n n k n k
k
i m i m i
, , 1
0
n
n k n k
k
m i m i
, , 1
0
n
n k n k
k
m i m i
= ,0 ,1 ,1 ,2 , , 1 , 1... 0n n n n n n n n n nm i m i m i m i m i m i m i
= ,0nm i
,0n ni m i (33)
Clearly all conditions of theorem are hold.
Hence
1
*
,0 1,
1 ;0 1; 1
log log ; 1n n
p
n nl i x O n n n m i
n n n
1
,0 1
1 ;0 11 1
log log ; 1n
n nO n n m i
n n n
(34)
This establishes corollary 2.
Remark The above result improves the result of G.
Das, R. N. Das and B. K. Ray [3] taking modified
partial sum *
nS x of conjugate Fourier series of
f x in place of nS x (see [8]).
References[1] Prem Chandra, On degree of
approximation of functions belonging to
the Lipschitz class, Nanta math. 8(1975),88-89
[2] G.Das, S.Pattanayak, Fundamentals of
mathematical analysis (1987), TataMCGraw-Hill Publishing companylimited, New Delhi
[3] G. Das, R.N. Das &B.K. Ray, Degree of
approximation of functions by their conjugate Fourier series in generalized
Holder metric, Journal of the Orissamathematical society, vol.17-20, 1998-2001, pp.61-74
[4] Shyam Lal, On degree of approximationof conjugate of a function belonging to
weighted , pW L t class by matrix
sumability means of conjugate series of a
Fourier series, Tamkang Journal of mathematics (2000)
[5] G.G.Lorentz(SyrocuseUniversity),
Approximation of functions (1948),(Athera series, Edwin Hewill, Editor)
[6] Kutbuddin Qureshi, On the degree of approximation of functions belonging to
the class Lip , p by means of
conjugate series, Indian J. Pure Applied,4(1981)
[7] E.C Titchmarsh (University of Oxford), Atheory of functions (1939), (Oxford
University Press, New York)
[8] A. Zygmund,Trigonometric series (1977)(Cambridge University Press, New York)vol.I