• Nie Znaleziono Wyników

On the rate of strong summability of Fourier-Chebyshev series

N/A
N/A
Protected

Academic year: 2021

Share "On the rate of strong summability of Fourier-Chebyshev series"

Copied!
9
0
0

Pełen tekst

(1)

Problemy Matematyczne 14 (1995), 27 - 36

On the rate of strong summability

of Fourier-Chebyshev series

W ło d zim ierz Łenski M aria Topolew ska

1

P r e lim in a r ie s .

Let us consider th e s y stem of Chebyshev polynomials

Tq(x) = 1, Tn( x ) = cos(n arccos x ) (n = 1 . 2 , . . . ) ,

which are ortogonal w ith th e wełght g(x) = (1 — x 2)~* on th e interval [—1,1]. Let / be a real-valued function continuous on [—1,1], an d let

OO

(1) S [ f ] { x ) - ^ ckTk{x)

k=0

b e its Fourier series w ith respect to the system t h a t is

(2) co = - [ f ( ł ) g ( t ) dt, ck = - f f ( t ) T k[ t ) g ( t ) d t ( k = 1 , 2 , . . . ) .

7r J- 1 7r J- 1

D en o te by S n (x; / ) th e n - th partial sum and by <r°(x) = c r " ( x ; / ) th e n -th C esaro (C, a ) - m e a n of o rd er a of the series (1). Intro d u ce th e m o d u lu s of c o n tin u ity of / defined by

u { 6 ) = u ( 6 , f ) = sup ( sup \ f ( y + h) - / ( t / ) | j ,

(2)

28 O N THE RATE OF S T R ON G S U MM AB IL I TY OF

an d th e best ap pro x im ation of / by polynomials Pn of th e degree a t m ost n given by

E n{ f ) = inf { sup | / ( x ) - P „ ( x ) | ) .

Pn

l -!<*<!

J

T ake into account a regular su m m ab ility m e th o d d e te rm in e d by a trian - gu lar m a tr ix | | a nfc/An || w ith a„jt > 0 and A n = I X = o Qnfc- W rite, for each n o n -n eg ativ e integer n,

1 i/p

(3) i2n ( x ; a ) p = 1 ( x ; / ) - / ( x ) | P| , p > 0.

T h e aim of this note is to estim a te th e ą u a n tity (3).

In considerations, th e suitable positive co n stan ts in d e p e n d e n t of n. f a n d

r are d en o ted by K j ( j = 1 , 2 , . . . ) , and

7 (j/) = m in (2‘/ — 2 , n ) , 71(1/) = 2t/~ \ 72( 1/) = m in (2,/+1 - 2 , n ) .

2

A u x ilia r y r e s u lts

We s t a r t w ith an analogue of th e B rudnyi and G opengauz th e o re m (T h . 4, [1], p. 892).

P r o p o sitio n .

Let

OO

(4) F ( z ) = ^ c kz k, 2 = r e “ , 0 < r < 1,

Jt=0

where ck are defined by (2). Then

(5) ^

f

“ 4 ~ dt unif ormly in x € [—tt; 7t ] .

J l - r t l

(3)

On THE R ATE OF STRONO SUMMA BI LI TY OF 29 1 rr = — j f ( c o s y ) { P { r , y + x ) + P ( r . y - .r)} dy + 1 f r + / f ( c o s v) { Q ( r iV + x ) - Q ( r , y - x )) dy = 9 i u (r i —x ) + u (r ’ x ) — i v (r, — x) + i v (r. x)} ,

w here u ( r . x ) . v ( r . x ) are the Poisson and th e Poisson co nju g ate integrals of

f { c o s y ) and P ( r , s ) . Q ( r , s ) are th eir kernels. respectivelv. T h u s

— F ( z ) ■ ( — d z 1 ’ \ d x O F ( z } d x 1 2 r d ( u (r, —a:) + u ( r . x ) ) , d ( v (r, — x) + v (r, x) ) d x — i d x

Therefore, it is sufficient to show th a t th e b o th term s of th e r ig h t-h a n d side of th e above relation are of th e order

Since we have 0 t 2uo{t)dt^j . l - , F s P { r ' s ) i s = 0 '

du

(r,

±x)

1

[*

dP{r,y±x)

di

=

i L f(c° sy)

di

dv =

=

±-

f

{/(cosy) - /(cosx)}

7r J - i r o y = ± -

f

{ /(c o s (f =F ■?)) ~ f { c o s x ) } - - -- - - - dt. 7 7 J - 7T o t

B}' th e m ean-value th eo rem , for arb itra ry real x , t and some 6 € (0 ,1 ), |/( c o s ( f + x ) ) — / ( c o s rr)| = \ f (cos x — t s in (x + 9t)) — f (cos x) \ <

< sup ( sup If ( y - h) - / ( y ) | | =

| / i | < l h L —1 < y » y — a < i J

(4)

30 O N T H E RATE OF STRONG S UMM AB IL I TY OF A nalogously, If (cos(t - a-)) - / ( c o s x )| < (tt + \ ) u f — ; / j . H ence d u (r, ± x ) d x sin 11 ( 1 —2r cos t + r 2)2 sin t dt = 7r J ((1 — r ) 2 + 4 r s i n 2( t / 2 ))2 t \ t

=

2(1

+ ; ) (1- rJ» r

~

^ " i '

^ ./i

\nj

((1—r}2-f-4r(£/5r)2)2

= 2,(» + l )(l-r

dt = A r g u m in g fu r th e r as in [5], pp. 150-151 we o btain du (r, ± .r) 8 x an d th is im m ed iately implies 1 d v ( r , ± x ) d x < A 3 f J l - T t 2ui{t)dt.

Now t h e desired assertion is evident.

R e m a r k 1. W h en x = arccos ?/, estim ate (5) holds uniform ly in y (E [—1,1],

L e m m a 1 . I f a > k > 0 and fc(l — a ) < 1, f/ien , ^ i/* (

6

) 1 2n . n + 1 Ś S u n i f o r m l y in x and n = 0 , 1 , 2 , . . . . < I<3

E

n + 1 *=0 a;

1

(5)

On T H E RATE OF STRONO S U MM AB IL I TY OF 31

P r o o f . Let us consider th e power series (4) and den o te by ~ "(y ) its m - th ( C . a ) - m e a n s for z = e ' arccosy. Easy calculation gives

i v a r c c o s y , u- 0 O -4“ « - ' ( ! / ) - < ( » ) ) = E - C - l ^ e " whence. for z = r ' v" (0 < r < 1), a £ / ) ~ T ° ( y ) ) z m = Z g , a r c c o s y F ' ( Z g i a r c c o s j / ) ( j _ m=0 F u rth e r, ta k in g a n u m b e r p € (1,2) such th a t p a > 1, by Hausdorff-Young in e ąu ality w ith q = we get

i 1.

' C O n } <> ( r * \ F f ( r f W p t a rc c o s y \ \ P ) P

- ki - r e - * r

M

.. 777 —0

In view of P roposition, th e right-hand side of th e above in e ąu ality does not exceed

7 i/p

K‘ r L i L rUt)dt)

\ p ((1 — r ) 2 + <d<p i22)°'p/2

S e ttin g 1 — r = 1/(2?? -f 1) and taking th e real p a r t we o b ta in our thesis (see [5], pp. 152-153).

L e m m a 2. Suppose that f o r each q > 1, the condition

( i 7(*d ) 1/,? "yW

(7) O E ( a , , ) * < A 6 ( 2 - ) £ a „ t

I, “ J r-=7i ((/)

^ 2"n - 1 . Then, for

i/p

P r o o f . P roceeding analogously as in [4] and using T h eo re m 3 from [6] we get th e desired result. holds f o r v 1 , 2 , whenever 2"" 1 < + P (0, oo), ' / M*/) \ - 1 7(p) ) i

E H

5 3 I M * ; / ) - / ( x ) | p , V = 7 l ( " ) / k=~ti (v) J

(6)

32 ON T H E R AT E OF S TR ONG S U MM AB IL I TY OF

3

M a in r e s u lts

Now we present two a p p ro x im atio n theorem s which are analogues of some L eindler's results ([2], [3], [4]).

T h e o r e m 1. Under the assumpti ons o f L c m m a 2, we have

j , yn / ~A,A \ 'i X/ P

R n { x ] l ) p < ^ £ | Y . Qnfc) (£'-yi(^)(/))P | f o r n = 0 ,1 ,2 ,

An u = l \fc=7)M

P r o o f . T h e above es tim a te is a consequence of L em m a 2.

In p a r tic u ła r if p — 1, a n* — A„Zl, & > 0 th e condition (7) holds. T h u s we o b ta in

C o r o l l a r y . Under the a ss umpti on 8 > 0,

-Je f l A n - l I5* (x ; / ) - f ( x ) I < K 9 — ~ T Y E ^ f ) f o r n = 0 . 1 , 2 , . . . .

An k=o n -r i k=0

P assin g to th e m ore generał m eans (3), let us p u t 1 J L / 1

y

K") =

-v k=o ^ + 1

T h e o r e m 2. Suppose that a > 1/2, p > 0, p ( l — a ) < 1 and that

f o r each q > 1, whenener 2m < n + 1 < 2m+1 — 1. Then

1 A t/p

(7)

On T H E R AT E OF STRONC, S I M M A BI LITY OF P r o o f . Clearly, { R n { x : a ) ) p < S 2' ( 7 - Ż k r ' ( i ) - < ( • ’■)[ + E - /< * ) ! ’ I !/=0 !/=0

= ^{E, + E

j-C h o o se two n u m b e rs p' ,q' > 1 such th a t pp'{ 1 — a ) < 1 and p + ^ = 1. view of e s tim a te (6) and condition (8).

72(0

H

Y1

q"" <

-*1" ;=o W=2' - i < 72(') PP < 72(0 < i - E B »j 5? 1 m / 1 E K “ - ‘w - < ( x ) -*1" /=0 i/=2( — 1 pp <

A" ^ ° '

" (2/— E

-

2

)jp <

< 1=0 (2A'3)p lri /=o £ 2 . T T T T T w - J < ( 2 A 3)p A i 0 r T T X > ( i O ) , 18 w w h e re #„,/ = ( a £ (^ + 1)* 1 U/=2'-l

(8)

34 On T H E R ATE OF STRONO S U MM A BI L I T Y OF r- m | 72 (0 1 2 e | 2 ((s- ,) Z < i / =2' - l A 1=0 ^ h 12

j

y ^ a n v ( lT ? ( , / ) ) PS 1 / r r r - 1 y y / >vp -

x S i £ ^ T T ) ^ )

S A . O A . - ^ E M w r

S u m m in g up these e s tim ates we o btain (9).

R e m a r k 2 . C ondition (8) is satisfied in th e special case when a nk = A SnZ\

an d fi > 0.

Indeed, observing th a t

*?(&)< <p{k') when k > k' an d

tp{21 - 1) < 2v?(Jfc) when 2' - 1 < A- < 2,+1 - 2

an d also th e fact th a t a nf: = A„Z\ (/? > 0) satisfy (7), we o bta in

^ ( A g : " ) 9 ( ^ ( ^ ) ) qp | 1/9 ^ 1—0 l i / = 2 ' - l

<

2 £ | M 2' - l ) f . 2' ^

Z ( A t l ) j

" ' U

m ( 72 (0 \ < 2 . A '„ t , M * - 1))” E < - ) < ^=0 \ u=2l —1 / m / 72(0 \ < 4 ■a' „ e e = i'=0 \i/=2! — 1 ) = 4 • A '13 £ A f c i (*(<'))” = i / = 0 /'[?] » \ = 4■ a', 3 E + E U S : l M O ) p < 1"=° «/=[Sl + l /

(9)

On T H E RATE OE S TRONG SIWIMABILITY OF 1Ź A i-i i / = 0 (n + 1)3- 1 D ^ + U - U ^ I + i J J ' Z < - l \ <

]+1

v

n

0-1

< /v15(t7 + 1)'3 f - ^ - r E M ^ r + i ^ D T

\n

+

1

tZo

)

R e fe r e n c e s

[1] BpyjjHbiH K ) .A ., r o n e H r a y 3 H .E . . 06o6w,eHue odn ou meopeMu.

X a p d u u JIu mmA beyd a. MaTeM. CSopHHK 52 (1960), 891-894.

[2] L eindler L., On stronę approiimation o f Fourier seińes, P eriodica M ath. H ung., Vol. 1 (2), (1971), 157-162.

[3] L eindler L., On the stronę approiimation o f orthoęonal series, A c ta Sci. M a th ., T. 37, Fasc. 1-2 (1975), 87-94.

[4] Leindler L., Uber dit Ap p ro ii ma t io n in starken S i n n e , A c ta M ath . Acad. Sci. H ung., T. 16. Fasc. 1-2 (1965). 255-262.

[5] Lenski W . , Generalizations o f iwo Leindler theorems, F unctiones e t Ap-

p r o x im a tio III. (1976), 149-155.

[6] R em p u lsk a L., Twierdzenia aproksymacyjne dla szereęów Fouriera-Czeby-

s ze wa , Fasciuli M a th e m a tic i Nr 5 (74), (1970), 63-69.

Streszczenie

O rzędzie m ocnej sumowalności szeregów Fouriera-Czebyszewa

U zyskano oszacowania uogólnionych mocnych dewiacji sum częściowych i (C . a ) - ś r e d n ic h szeregów Fouriera-Czebyszewa funkcji ciągłych od tych fu n ­ kcji. B łęd y oszacowań wyrażone zostały w term in ach najlepszych przybliżeń i m o d u łó w ciągłości. Udowodnione twierdzenia stanow ią uogólnione o d p o ­ w iedniki wyników L. Leindlera.

P E D A G O G I C A L U N I V E R S I T Y P E D A G O G I C A L U N I V E R S I T Y I N S T I T U T E O F M A T H E M A T IC S I N S T I T U T E O F M A T H E M A T C S

Zielona Góra 65-069 Bydgoszcz 85-064

Plac Słowiański 9 Chodkiewicza 30

Cytaty

Powiązane dokumenty

ANNALES SOCIETATIS MATHEMATICAE POLONAE Series I: COMMENTATIONES MATHEMATICAE XXV (1985) ROCZNIKI POLSKIEGO TOWARZYSTWA MATEMATYCZNEGO.. Séria I: PRACE MATEMATYCZNE XXV

Saxena [6 ] and Dwivedi [1] studied the uniform harmonie summability of Fourier series and uniform harmonic summability of Legendre series respectively... of

Patel, Reader in Applied Mathematics, Faculty of Technology and Engineering for his encouragement and valuable suggestion for the preparation of this

Julian Musielak for his kind criticism during the preparation of this paper2. Remarks on Cesaro

Siddiqi, On the harmonic summability of Fourier series, Proc.. Singh, Nôrlund summability of Fourier series and its conjugate series,

Sharma for his valuable suggestions during the preparation of this

This paper presents a method for training a Fourier series neural network on the basis of the multidimensional discrete Fourier transform.. The proposed method is characterized by

Rivlin, Chebyshev Polynomials: From Approximation Theory to Algebra and Number Theory, second edition, Wiley Interscience 1990. Received 30 April 2007 Revised 24