Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)

We point out that if the Hardy–Littlewood maximal operator is bounded on the space Lp(t)(ℝ), 1 < a ≤ p(t) ≤ b < ∞, t ∈ ℝ, then the well-known characterization of the spaces Lp(ℝ), 1 < p < ∞, by the Littlewood–Paley theory extends to the space L p(t)(ℝ). We show that, for n > 1 , the L...

Full description

Saved in:
Bibliographic Details
Published in:Український математичний журнал
Date:2008
Main Author: Kopaliani, T.S.
Format: Article
Language:English
Published: Інститут математики НАН України 2008
Subjects:
Online Access:https://nasplib.isofts.kiev.ua/handle/123456789/164982
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ) / T.S. Kopaliani // Український математичний журнал. — 2008. — Т. 60, № 12. — С. 1709–1715. — Бібліогр.: 13 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-164982
record_format dspace
spelling Kopaliani, T.S.
2020-02-11T12:03:28Z
2020-02-11T12:03:28Z
2008
Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ) / T.S. Kopaliani // Український математичний журнал. — 2008. — Т. 60, № 12. — С. 1709–1715. — Бібліогр.: 13 назв. — англ.
1027-3190
https://nasplib.isofts.kiev.ua/handle/123456789/164982
517.5
We point out that if the Hardy–Littlewood maximal operator is bounded on the space Lp(t)(ℝ), 1 < a ≤ p(t) ≤ b < ∞, t ∈ ℝ, then the well-known characterization of the spaces Lp(ℝ), 1 < p < ∞, by the Littlewood–Paley theory extends to the space L p(t)(ℝ). We show that, for n > 1 , the Littlewood–Paley operator is bounded on Lp(t) (ℝⁿ), 1 < a ≤ p(t) ≤ b < ∞, t ∈ ℝⁿ, if and only if p(t) = const.
Встановлено, що коли максимальний оператор Харді - Літтлвуда обмежений на просторі Lp(t)(Rⁿ), 1<a ≤ p(t) ≤ b < ∞,t∈R, добре відома характеризація просторів Lp(t)(Rⁿ),1<p<∞ теорією Літтлвуда - Пелі поширюється на простір Lp(t)(Rⁿ). Показано, що у випадку n>1, оператор Літтлвуда - Пелі обмежений на Lp(t)(Rⁿ),1 < a ≤ p(t) ≤ b<∞,t ∈ R, тоді і тільки тоді, коли p(t)= const.
The author was supported by grant GNSF / STO 7 / 3-171.
en
Інститут математики НАН України
Український математичний журнал
Короткі повідомлення
Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
Tеорема Литтлвуда - Пелі про простори Lp(t)(ℝⁿ)
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
spellingShingle Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
Kopaliani, T.S.
Короткі повідомлення
title_short Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
title_full Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
title_fullStr Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
title_full_unstemmed Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ)
title_sort littlewood–paley theorem on spaces lp(t)(ℝⁿ)
author Kopaliani, T.S.
author_facet Kopaliani, T.S.
topic Короткі повідомлення
topic_facet Короткі повідомлення
publishDate 2008
language English
container_title Український математичний журнал
publisher Інститут математики НАН України
format Article
title_alt Tеорема Литтлвуда - Пелі про простори Lp(t)(ℝⁿ)
description We point out that if the Hardy–Littlewood maximal operator is bounded on the space Lp(t)(ℝ), 1 < a ≤ p(t) ≤ b < ∞, t ∈ ℝ, then the well-known characterization of the spaces Lp(ℝ), 1 < p < ∞, by the Littlewood–Paley theory extends to the space L p(t)(ℝ). We show that, for n > 1 , the Littlewood–Paley operator is bounded on Lp(t) (ℝⁿ), 1 < a ≤ p(t) ≤ b < ∞, t ∈ ℝⁿ, if and only if p(t) = const. Встановлено, що коли максимальний оператор Харді - Літтлвуда обмежений на просторі Lp(t)(Rⁿ), 1<a ≤ p(t) ≤ b < ∞,t∈R, добре відома характеризація просторів Lp(t)(Rⁿ),1<p<∞ теорією Літтлвуда - Пелі поширюється на простір Lp(t)(Rⁿ). Показано, що у випадку n>1, оператор Літтлвуда - Пелі обмежений на Lp(t)(Rⁿ),1 < a ≤ p(t) ≤ b<∞,t ∈ R, тоді і тільки тоді, коли p(t)= const.
issn 1027-3190
url https://nasplib.isofts.kiev.ua/handle/123456789/164982
citation_txt Littlewood–Paley theorem on spaces Lp(t)(ℝⁿ) / T.S. Kopaliani // Український математичний журнал. — 2008. — Т. 60, № 12. — С. 1709–1715. — Бібліогр.: 13 назв. — англ.
work_keys_str_mv AT kopalianits littlewoodpaleytheoremonspaceslptrn
AT kopalianits teoremalittlvudapelíproprostorilptrn
first_indexed 2025-11-27T02:33:03Z
last_indexed 2025-11-27T02:33:03Z
_version_ 1850794619223146496
fulltext K O R O T K I P O V I D O M L E N N Q UDC 517.5 T. S. Kopaliani (Tbilisi Univ., Georgia) LITTLEWOOD – PALEY THEOREM ON Lp(t) n( )R SPACES∗∗∗∗ TEOREMA LYTTLVUDA – PELI PRO PROSTORY Lp(t) n( )R We point out that when the Hardy – Littlewood maximal operator is bounded on the space Lp t( ) ( )R , 1 < a ≤ p t( ) ≤ b < ∞ , t ∈R , the well-known characterization of spaces Lp( )R , 1 < p < ∞ , by the Littlewood – Paley theory extends to the space Lp t( ) ( )R . We show that if n > 1, the Littlewood – Paley operator is bounded on Lp t n( ) ( )R , 1 < a ≤ p t( ) ≤ b < ∞ , t n∈R , if and only if p t( ) = const . Vstanovleno, wo koly maksymal\nyj operator Xardi – Littlvuda obmeΩenyj na prostori Lp t( ) ( )R , 1 < a ≤ p t( ) ≤ b < ∞ , t ∈R , dobre vidoma xarakteryzaciq prostoriv Lp( )R , 1 < p < < ∞, teori[g Littlvuda – Peli poßyrg[t\sq na prostir Lp t( ) ( )R . Pokazano, wo u vypadku n > 1 operator Littlvuda – Peli obmeΩenyj na Lp t n( ) ( )R , 1 < a ≤ p t( ) ≤ b < ∞ , t n∈R , todi i til\ky todi, koly p t( ) = const . 1. Introduction. Let m be a bounded function on Rn. The operator T defined by the Fourier transform equation ( )̂ ( )Tf x = m x f x( ) ˆ( ), x ∈ R n, is called a multiplier operator with multiplier m . Let ρ be an ( n -dimensional) rectangle and χρ the cha- racteristic function of ρ . The operator Sρ having multiplier m = ρ and defined by the equation ( )̂ ( )S f xρ = χρ( ) ˆ( )x f x , x ∈ R n, is called a partial sum operator. Let a collection of disjoint rectangles ∆ = { }ρ be a decomposition of R n ( i.e., ∪ρ∈∆ = Rn) . Given a function f in the Schwartz class S ( )R n , define G f x( )( ) = ρ ρ ∈ ∑       ∆ S f x( ) / 2 1 2 , x ∈ R n. Let { }nk k = −∞ +∞ , nk > 0, k ∈ Z , be a lacunary sequence ( i.e., there is an a > 1 such that n nk k+1/ ≥ a for all k ) . Let ∆ be the collection of all intervals of the form [ ],n nk k+1 and [ ],− +n nk k1 , k ∈ Z . Then ∆ is called a lacunary decomposition of R . When nk = 2 k, k ∈ Z, the resulting ∆ is called the dyadic decomposition of R . ∗ The author was supported by grant GNSF / STO 7 / 3-171. © T. S. KOPALIANI, 2008 ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 1709 1710 T. S. KOPALIANI Let ∆i , i = 1, 2, … , n , be n lacunary decomposition of R . Let ∆ be the collection of the intervals of the form ρ = ρ ρ ρ1 2× × … × n where ρi i∈∆ . Then ∆ is called a lacunary decomposition of Rn. The important feature of the classical Littlewood – Paley theory is that a characteri- zation of the spaces Lp n( )R , 1 < p < ∞ . It is well known (see [1, 2]) that if ∆ is a lacunary decomposition of Rn then G f p( ) is equivalent to f p for 1 < p < < ∞ ; i.e., there are constants A and B such that A f p ≤ G f p( ) ≤ f p . The purpose of this paper is to obtain analogously characterizations of variable ex- ponent Lebesgue spaces Lp t n( )( )R . Given a measurable functions p n( ) : [ , )⋅ → ∞R 1 , Lp t n( )( )R denotes the set of measurable functions f on Rn such that for some λ > 0 R n f x dx p x ∫     ( ) ( ) λ < ∞ . This set becomes a Banach function space when equipped with the norm f p t( ) = inf : ( ) ( ) λ λ >     ≤     ∫0 1f x dx p x . Given a locally integrable function f, we define the Hardy – Littlewood maximal function M f by M f ( x ) = sup ( ) x Q Q Q f y dy ∈ ∫1 , where the supremum is taken over all cubes containing x with sides parallel to the co- ordinate axes. For conciseness, define P ( )R n to be the set of measurable function p n( ) : [ , )⋅ → ∞R 1 such that 1 < a ≤ p ( t ) ≤ b < ∞ : t ∈ R n. Let B ( )R n be the set of p n( ) ( )⋅ ∈P R such that M is bounded on Lp t n( )( )R . Con- ditions for the boundedness of the Hardy – Littlewood maximal operator on spaces Lp t n( )( )R have been studied in [3 – 8]. Diening [8] studied the necessary and suffici- ent conditions in terms of the conjugate exponent ′ ⋅p ( ), ( 1 1/ /( ) ( )p t p t+ ′ = 1, t(∈ ∈ R n ) . He has proved that p n( ) ( )⋅ ∈B R is equivalent to ′ ⋅ ∈p n( ) ( )B R , he also proved that if p n( ) ( )⋅ ∈B R then p q n( ) ( )/⋅ ∈B R for some q > 1. In harmonic analysis a fundamental operator is the Hardy – Littlewood maximal operator. In many applications a crutial step has been to show that operator M is bo- unded on a variable Lp space. Cruz-Uribe, Fiorenza, Martell and Perez [4] have show- ed that many classical operators in harmonic analysis such as singular integrals, com- mutators and fractional integrals are bounded on the variable Lebesgue space Lp t n( )( )R whenever the Hardy – Littlewood maximal operator is bounded on Lp t n( )( )R . If we consider, instead, the strong maximal operator MR defined by M f xR ( )( ) = sup ( ) x R R R f x dx ∈ ∫1 ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 LITTLEWOOD – PALEY THEOREM … 1711 where R is any rectangle in Rn , n > 1, with sides parallel to the coordinate axes then the situation is different. For the strong Hardy – Littlewood maximal operator MR we prove following theorem. Theorem 1. Let 1 ≤ p ( t ) ≤ b < ∞ , t ∈ R n. The strong Hardy – Littlewood maximal operator MR is bounded on Lp n( )( )⋅ R space if and only if p ( t ) = = const = p and p > 1. For function f L n∈ ( )R , the expression H f ( x ) = R n i n k kx y f y dy∫ ∏ = −1 1 ( ) is said to be n-dimensional ( )n > 1 Hilbert operator. Analogously we may prove following theorem. Theorem 2. Let 1 ≤ p ( t ) ≤ b < ∞ , t ∈ R n. Then n-dimensional Hilbert ope- rator ( )n > 1 is bounded on Lp n( )( )⋅ R space if and only if p ( t ) = const = p and p > 1. We prove following Littlewood – Paley type characterization of Lp t n( )( )R space. Theorem 3. 1. Let ∆ be a lacunary decomposition of R and p( ) ( )⋅ ∈B R . Then there are constants c, C > 0 such that for all f Lp t∈ ( )( )R c f p t( ) ≤ G f p t( ) ( ) ≤ C f p t( ). (1) 2. Let ∆ be the dyadic decomposition of Rn, n > 1. If p( )⋅ ≠ const then operator G is not bounded on Lp t n( )( )R . 2. Proof of theorems. Proof of Theorem 1. According to Jessen, Marcinkiewicz and Zygmund [9] MR is bounded on all the Lp, p > 1, spaces and first part of Theorem 1 is trivial. Let MR is bounded on Lp n( )( )⋅ R . Virtue of interpolation theorem (see [10]), we have MR is bounded on Lp( )/⋅ θ = [ ]( )( ), ( )L Lp n n⋅ ∞ R R θ, 0 < θ < 1, and without restriction of generality we may assume that 1 < inf ( ) R n p t . Let 1 1/ /( ) ( )p t p t+ ′ = = 1, t(∈ R n . Note that sup ( ) ( ) R R p t R p tR 1 χ χ ′ < ∞ (2) condition is necessary for boundedness of MR on Lp t n( )( )R (see proof below). We will give the proof of second part of Theorem 1 for the case n = 2 for simpli- city, since the same argument holds when n > 2. Let inf ( ) R 2 p t < sup ( ) R 2 p t . By Luzin’s theorem we can construct pairwise dis- joint family of set Fi with the following condition: 1) R 2 \ ∪ Fi = 0, 2) functions p Fi: → R are continuous, 3) for every fixed i all points of Fi are points of densi- ty with respect to basis R . Note that, we can find pair of points (( , ), ( , ))x y x y0 1 0 2 - or (( , ), ( , ))x y x y1 0 2 0 -type from ∪ Fi such that p x y( , )0 1 ≠ p x y( , )0 2 or p x y( , )1 0 ≠ p x y( , )2 0 . Without loss of generality, we may suppose that this pair is (( , ), ( , ))x y x y0 1 0 2 ; ( , )x y F0 1 1∈ , ( , )x y F0 2 2∈ and y1 < y2 . ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 1712 T. S. KOPALIANI Let 0 < ε < 1 be fixed. We may find δ > 0 such that for any rectangles Q x y1 0 1' ( , ) , Q x y2 0 1' ( , ) with diameters loss than δ the following inequalities are valid: Q F1 1∩ > ( )1 1− ε Q , Q F2 2∩ > ( )1 2− ε Q , (3) pQ1 = sup ( , ) Q F p x y 1 1∩ < c1 < c2 < inf ( , ) Q F p x y 2 2∩ = pQ2 (4) for some constant c1, c2. Let Q1,τ , Q2,τ are rectangles with properties (3), (4) of the form ( , )x x0 0− +τ τ × × ( , )a b , ( , ) ( , )x x c d0 0− + ×τ τ , where a < b < c < d. We have continuously embedding L Q L Qp t pQ( ) , ,( ) ( )2 2 2 τ τO and L Qp t′( ) ,( )1 τ O O L Q pQ′ 1 1( ),τ , where 1 1 1 1/ /′ +p pQ Q = 1 (see for example [11]). For rectangle Qτ = = ( , ) ( , )x x a d0 0− + ×τ τ we have Aτ = 1 Q Q p t Q p tτ χ χ τ τ( ) ( )′ ≥ 1 2 2 2 1 1τ χ χ τ τ( ) , ,( ) ( )d a Q F p t Q F p t− ′∩ ∩ ≥ ≥ C d a d c b a p pQ Q 2 2 2 1 1 1 2 1 τ τ τ ( ) ( ( )) ( ( )) / / − − − − . Note that if τ → 0 ( a, b, c, d is fixed ) Aτ → ∞ and consequently (2) is not va- lid. This completes the proof. Proof of Theorem 3. The inequalities (1) are consequence of the extrapolation theorem given by Cruz-Uribe, Fiorenza, Martell and Perez [4] and the weighted norm inequalities for G ( f ) function given by Kurtz [12]. We describe this results. Let p− = ess inf{ ( ) : }p x x ∈R . By a weight we mean a nonnegative, locally in- tegrable function ω . When 1 < p < ∞ , we say ω ∈Ap if for every interval Q 1 1 1 1 Q x dx Q x dx Q Q p p ∫ ∫ − ′ −     ω ω( ) ( ) ≤ C < ∞ . The infimum over the constants on the right-hand side of the last inequality we de- note by Ap,ω . By F will denote a family of ordered pairs of nonnegative, measurable functions ( f , g ) . We say that an inequality R ∫ f x x dxp( ) ( )0 ω ≤ C g x x dxp R ∫ ( ) ( )0 ω , 0 < p0 < ∞ , (5) holds for any ( f , g ) ∈ F and ω ∈ Aq (for some q, 1 < q < ∞ ) if it holds for any pair in F such that the left-hand side is finite, and the constant C depends only on p0 and the Aq,ω constant of ω . Theorem 4. Given a family F , assume that (5) holds for some 1 < p0 < ∞ , for every weight ω ∈ Ap0 and for all ( f , g ) ∈ F . Let p( ) ( )⋅ ∈P R be such that there exists 1 < p1 < p– , with ( / )( ) ( )p p⋅ ′ ∈1 B R . Then f p t( ) ≤ C g p t( ) for all ( f , g ) ∈ F such that f Lp t∈ ( )( )R . Theorem 5 [12]. Let ∆ be a a lacunary decomposition of R , 1 < p < ∞ , and ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 LITTLEWOOD – PALEY THEOREM … 1713 ω(∈ Ap . Then there exist constant c, C depending only on p , Ap,ω , and ∆ , such that c f x x dxp R ∫ ( ) ( )ω ≤ R ∫ ( ( )( )) ( )G f x x dxpω ≤ C f x x dxp R ∫ ( ) ( )ω . From assumption of Theorem 3 we get that there exists 1 < p1 < p– with ( / )( ) ( )p p⋅ ′ ∈1 B R (see [8]). Let Lcomp ∞ ( )R be the set of all bounded functions with compact support. From Theorems 4, 5 with the pairs ( ),W f f we get right side in- equality of (1) if f L∈ ∞ comp( )R . Note that Lcomp ∞ ( )R is dense in Lp t( )( )R (see [11]) and consequently this inequality is also valid for all f Lp t∈ ( )( )R . Analogously we ob- tain left side inequality of (1). Let n > 1. Fix a rectangle R = I I In1 2× × … × and let f be positive on R and 0 elsewhere function. Let kj be the greatest integer such that 2 kj ≤ ( )4 1n Ij − and ρ be the dyadic rectangle [ ],2 21 1 1k k + × … × [ ],2 2 1k kn n + . Note that (see [12, p. 246]) for all x ∈ R S f xρ ( ) ≥ C R f x dx R ∫ ( ) . Let the operator G is bounded on Lp t n( )( )R . Then for some constant C we have 1 Q f x dx R R p t∫ ( ) ( )χ ≤ C f p t( ). (6) Note that ( ( ))( )Lp t n R ∗ is isomorphic to the space Lp t n′( )( )R , where 1/ ( )p t + + 1/ ( )′p t = 1, t(∈ R n (see [11]). Therefore, for all rectangle R , from (6) we get condition (2). We use Theorem 1 to obtain the desired result. 3. Applications. We now consider applications of Theorem 3. In [7] is proved following theorem. Theorem 6. Let p( ) ( )⋅ ∈P R and exponent p( )⋅ is constant outside some large ball. Then operator M is bounded on Lp t( )( )R if and only if (2) fulfilled for intervals. The estimate (2) is necessary for boundedness of operator M in Lp t( )( )R . Com- bining the Littlewood – Paley type characterization of Lp t( )( )R space (Theorem 3) with the previous theorem we can obtain the following corollary. Corollary. Let p( ) ( )⋅ ∈P R and exponent p( )⋅ is constant outside some lar- ge ball. Let ∆ be the dyadic decomposition of R . The following are equivalent: 1) p( ) ( )⋅ ∈B R ; 2) there are constants c, C > 0 such that for all f Lp t∈ ( )( )R c f p t( ) ≤ G f p t( ) ( ) ≤ C f p t( ). Let { }fk be a sequence of functions defined on R . By f Lkk p t∑ ∈ ( )( )R we me- an the partial sums fk N 1∑ converge in Lp t( )( )R . We now will generalize Theorem 6 of Stein [13]. ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 1714 T. S. KOPALIANI Theorem 7. Let p( ) ( )⋅ ∈B R , and Sk be any collection of lacunary partial sum operators. Then f Lp t∈ ( )( )R if and only if εk kk S f∑ converges in Lp t( )( )R for any sequence { }εk l∈ ∞ . Moreover , f p t( ) is equivalent to sup { } ( )ε ε k l k kk p t S f∞ = ∑1 . Proof. Let p( ) ( )⋅ ∈B R and Sk be any collection of lacunary partial sum. For f Lp t∈ ( )( )R we have S f Lkk p t2 1 2∑( ) ∈ / ( )( )R . Note that if { }εk l∈ ∞ then εk kk p tS f L2 1 2∑( ) ∈ / ( )( )R and εk k k S f 2 1 2 ∑    / ≤ { } / εk l k k S f∞ ∑    2 1 2 . If N > M using Theorem 3, εk k M N p t S f + ∑ 1 ( ) ≤ C S fk l k M N p t { } / ( ) ε ∞ + ∑     2 1 1 2 which implies εk k N S f 1 1 ∑{ }∞ is Cauchy in Lp t( )( )R . From this follows εk k k p t S f∑ ( ) ≤ c S fk l k k p t { } / ( ) ε ∞ ∑    2 1 2 . Assume that Sk be any collection of lacunary partial sum operators and εk kk S f∑ ∈ Lp t( )( )R for all { }εk l∈ ∞ . We will prove that S f Lkk p t2 1 2∑( ) ∈ / ( )( )R and there exists a constant c > 0 independent of f such that S fk k p t 2 1 2 ∑    / ( ) ≤ c S f k l k k k p t sup { } ( )ε ε ∞ = ∑ 1 . (7) First we will prove that M = sup { } ( )ε ε k l k kk p t S f∞ = ∑1 is finite. Indeed, consi- der the collection of maps G l LN p t: ( )( )∞ →{ }R defined by GN k( ){ }ε = εk kk N S f=∑ 1 . Let G = G∞ . Each GN is continuous and by assumption GN k( ){ }ε converges to G k( ){ }ε in Lp t( )( )R for each { }εk l∈ ∞ . Therefore GN k p t N ( ){ } ( )ε{ } = ∞ 1 is bounded for each { }εk l∈ ∞ . By the principle of uniform boundedness, there exists a constant c > 0 such that GN ≤ c for all N. It follows that G ≤ c . To proof of (7) will use Khinchine’s inequality for Rademacher series. Let r tk( ) = = sgn(sin )2m tπ , m = 0, 1, 2, … , be the Rademacher functions, and set f = = a rm m0 ∞∑ . Then there are constants Bp and Cp such that for 0 < p < ∞ B f t dtp p p 0 1 1 ∫      ( ) / ≤ m ma = ∞ ∑   0 2 1 2/ ≤ C f t dtp p p 0 1 1 ∫      ( ) / (8) (see [2]). Let εk = r tk( ) for 0 ≤ t < 1. Then { }εk l∞ = 1 and ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12 LITTLEWOOD – PALEY THEOREM … 1715 M ≥ r x S fk k k p t ( ) ( ) ∑ . Using (8) for p = 1 and Fubini’s theorem we have S fk k p t 2 1 2 ∑    / ( ) ≤ C r x S f dx k k k p t 1 0 1 ∫ ∑ ( ) ( ) ≤ ≤ cC r x S f z dx g z dz g k k kp t 1 1 0 1 sup ( ) ( ) ( ) ( )′ ≤ ∫ ∫ ∑       R = = cC r x S f z g z dz dx g k k kp t 1 1 0 1 sup ( ) ( ) ( ) ( )′ ≤ ∫ ∫ ∑        R ≤ ≤ cC r x S f z g z dz dx g k k kq t 1 0 1 1 ∫ ∫ ∑ ≤          sup ( ) ( ) ( ) ( ) R ≤ ≤ cC r x S f dxk k k p t 1 0 1 ∫ ∑ ( ) ( ) ≤ cC M1 , which proves (7). This completes the proof of Theorem 7. 1. Stein E. M. Singular integrals and differentiability properties of functions. – Princeton: Princeton Univ. Press, 1970. 2. Zygmund A. Trigonometric series. – 2nd ed. – London; New York: Cambridge Univ. Press, 1959. 3. Diening L. Maximal function generalized Lebesgue spaces Lp( )⋅ // Math. Inequal. Appl. – 2004. – 7. – P. 245 – 254. 4. Cruz-Uribe D., Fiorenza A., Martell J. M., Perez C. The boundedness of classical operators on variable Lp spaces // Ann. Acad. sci. fenn. math. – 2006. – 31. – P. 239 – 264. 5. Cruz-Uribe D., Fiorenza A., Neugebauer C. J. The maximal function on variable Lp spaces // Ibid. – 2003. – 28. – P. 223 – 238; 2004. – 29. – P. 247 – 249. 6. Nekvinda A. Hardy – Littlewood maximal operator on L Rp x n( ) ( ) // Math. Inequal. Appl. – 2004. – 7. – P. 255 – 266. 7. Kopaliani T. S. Infinitesimal convolution and Muckenhoupt Ap( )⋅ condition in variable Lp spaces // Arch. Math. – 2007. – 89, # 2. – P. 185 – 192. 8. Diening L. Maximal function on Orlicz – Musielak spaces and generalized Lebesgue spaces // Bull. sci. math. – 2005. – 129. – P. 657 – 700. 9. Jessen B., Marcinkiewicz J., Zygmund A. Note on the differentiability of multiple integrals // Fund. Math. – 1935. – 25. – P. 217. 10. Diening L., Hästö P., Nekvinda A. Open problems in variable exponent Lebesgue and Sobolev spaces // FSDONA 04 Proc. (Milovy, Czech. Rep., 2004) / Eds Drabek and Rakosnik. – P. 38 – 58. 11. Kováčik O., Rákosnik J. On spaces Lp t( ) and W k p x, ( ) // Czech. Math. J. – 1991. – 41, # 4. – P. 592 – 618. 12. Kurtz D. S. Littlewood – Paley and multiplier theorems on weighted Lp spaces // Trans. Amer. Math. Soc. – 1980. – 259. – P. 235 – 254. 13. Stein E. M. Classes H p , multiplicateurs et fonctions de Littlewood – Paley // C. r. Acad. sci. – 1966. – 263. – P. 716 – 719, 780 – 781. Received 15.10.07 ISSN 1027-3190. Ukr. mat. Ωurn., 2008, t. 60, # 12