VILENKIN-FOURIER SERIES IN VARIABLE LEBESGUE SPACES DAVITI ADAMADZE AND TENGIZ KOPALIANI

2025-05-06 0 0 416.51KB 15 页 10玖币
侵权投诉
VILENKIN-FOURIER SERIES IN VARIABLE LEBESGUE
SPACES
DAVITI ADAMADZE AND TENGIZ KOPALIANI
Abstract. Let Snfdenote the nth partial sum of the Vilenkin-Fourier series
of a function fL1(G). For 1 < pp+<, we characterize all exponents
p(·) for which the convergence of Snfto fin Lp(·)(G) holds whenever f
Lp(·)(G).
1. Introduction
Let {pi}i0be a sequence of integers with pi2. Define G= Π
i=0Zpias the
direct product of cyclic groups of order pi, and let µbe the Haar measure on G,
normalized such that µ(G) = 1. Each element of Gcan be represented as a sequence
{xi}with 0 xi< pi. Define m0= 1 and mk= Πk1
i=0 pifor k= 1,2, . . . .
There exists a well-known and natural measure-preserving identification between
the group Gand the closed interval [0,1]. This identification is established by as-
sociating each sequence {xi} ∈ G, where 0 xi< pi, with the point P
i=0 xim1
i+1.
Disregarding the countable set of pi-rationals, this mapping is one-to-one, onto, and
measure-preserving.
For each x={xi} ∈ G, define ϕk(x) = exp(2πixk/pk), k = 0,1, .... The set
{ψn}of characters of Gconsists of all finite product of ϕk,which we enumerate
in the following manner. Express each nonnegative integer nas a finite sum n=
P
i=0 αkmk, with 0 αk< pk,and define ψn= Π
i=0ϕαk
k.The functions ψnform a
complete orthonormal system on G. For the case pi= 2, i = 0,1, ..., G is the dyadic
group, ϕkare Rademacher functions and ψnare Walsh functions. In general, the
system {ψn}is a realization of the multiplicative Vilenkin system. In this paper,
there is no restriction on the orders {pi}.
For fL1(G), let Snf, n = 0,1, ..., be the nth partial sum of the Vilenkin-
Fourier series of f. When the orders piof cyclic groups are bounded Watari [19]
showed that for fLp(G),1<p<,
lim
n→∞ ZG
|Snff|p= 0.
Young [17], Schipp [14] and Simon [15] showed independently that results concern-
ing mean convergence of partial sums of the Vilenkin-Fourier series are still valid
even if the orders piare unbounded.
Date: 29.01.2022.
2020 Mathematics Subject Classification. 42C10, 42B25,46E30.
Key words and phrases. Vilenkin-Fourier series; maximal operator; variable exponent Lebesgue
space.
This work was supported by Shota Rustaveli National Science Foundation of Georgia FR-21-
12353.
1
arXiv:2210.11331v4 [math.FA] 17 Feb 2025
2 DAVITI ADAMADZE AND TENGIZ KOPALIANI
Let {Gk}be the sequence of subgroups of Gdefined by
G0=G, Gk= Πk1
i=0 {0} × Π
i=kZpi, k = 1,2, ....
On the closed interval [0,1], cosets of Gkare intervals of the form [jm1
k,(j+
1)m1
k], j = 0,1, ..., mk1.By Fwe denote the set of generalized intervals. This set
is the collection of all translations of intervals [0, jm1
k+1], k = 0,1, ... j = 1, ..., pk.
Note that a set Ibelongs to Fif (1) for some xGand k, I x+Gk,(ii) Iis a
union of cosets of Gk+1,and (iii) if we consider x+Gkas a circle, Iis an interval.
Let F1={G}. For k= 0,1, ..., let Fkbe the collection of all I∈ F such that Iis
a proper subset of a coset of Gk, and is a union of cosets of Gk+1.The collections
Fkare disjoint, and F=
k=1Fk.For I∈ F, we define the set 3I∈ F as follows.
If I=G, let 3I=G. For I∈ Fk, k = 0,1, ..., there is xGsuch that Ix+Gk.
If µ(I)µ(Gk)
3, let 3I=x+Gk. If µ(I)<µ(Gk)
3, consider x+Gkas a circle. Then
Iis an interval in this circle. Define 3I∈ Fkto be the interval in this circle which
contains Iat its center and has measure µ(3I) = 3µ(I). In all cases, for I∈ F,
µ(3I)3µ(I).
We say that wis a weight function on Gif wis measurable and 0 < w(x)<
a.e. Gosselin [7] (case supipi<) and Young [18] (no restriction on the orders
pi) characterized all weight functions wsuch that if fLp
w(G),1< p < , Snf
converges to fin Lp
w(G). Here Lp
w(G) denotes the space of measurable functions
on Gsuch that fp,w = (RG|f|pw)1/p <.
Definition 1.1. (see [18]) (i) We say that wsatisfies Ap(G) condition, 1 <p<,
if
(1.1) [w]Ap= sup
I∈F 1
µ(I)ZI
w dµ1
µ(I)ZI
w1/(p1) p1
<.
(ii) We say that wsatisfies A1(G) condition if
[w]A1= sup
I∈F
1
µ(I)ZI
w(essinfIw(x))1<.
For the case where the orders of cyclic groups are bounded, Gosselin [7] defined
Ap(G) condition, as the one where (1.1) condition holds for all Ithat are cosets of
Gk, k = 0,1,2, .... For this case Apconditions, defined by Young and Gosselin, are
equivalent (see [18]).
Theorem 1.2. ([18]) Let wbe a weight function on G. For 1< p < , the
following statements are equivalent:
(i) wAp(G),
(ii) There is a constant C, depending only on wand p, such that for every
fLp
w(G), we have
ZG
|Snf|pwCZG
|f|pwdµ,
(iii) For every fLp
w(G), we have
lim
n→∞ ZG
|Snff|pw= 0.
In this paper we characterize all exponents p(·) such that if fLp(·)(G), then
partial sums Snfof the Vilenkin-Fourier series of fLp(·)(G) converge to fwith
Lp(·)-norm. Now we give a definition of variable Lebesgue space. Let p(·) : G
3
[1,) be a measurable function. The variable Lebesgue space Lp(·)(G) is the set
of all measurable functions fsuch that for some λ > 0,
ρp(·)(f) = ZG
(|f(x)|)p(x)dµ < .
Lp(·)(G) is a Banach function space equipped with the Luxemburg norm
fp(·)= inf{λ > 0 : ρp(·)(f)1}.
We use the notations p(I) = essinfxIp(x) and p+(I) = esssupxIp(x) where
IG. If I=Gwe simply use the following notation p, p+. The function p(·)
denotes the conjugate exponent function of p(·), i.e., 1/p(x)+1/p(x) = 1 (xG).
In this paper the constants C, c are absolute constants and may be different in
different contexts and χAdenotes the characteristic function of set A.
Very recently the convergence of partial sums of the Walsh-Fourier series in
Lp(·)([0,1)) space was investigated by Jiao et al. [8]. We denote by Clog
dthe set of
all functions p(·) : [0,1) [1), for which there exists a positive constant Csuch
that
|I|p(I)p+(I)C
for all dyadic intervals I= [k2n,(k+ 1)2n) (k, n N,0k < 2n), here |I|
denotes the Lebesgue measure of I. Note that this condition may be interpreted
as a dyadic version of log-H¨older continuity condition of p(·) (or on dyadic group).
The log-H¨older condition is a very common condition for solving various problems
of harmonic analysis in Lp(·)(Rn) (see [2], [5]).
Theorem 1.3. ([8]) Let p(·)Clog
dwith 1< pp+<.If fLp(·)([0,1)),
then for partial sums Snfof the Walsh-Fourier series of fLp(·)([0,1)) we have
sup
nN
Snfp(·)Cfp(·).
Since Walsh polynomials are dense in Lp(·)([0,1)), Theorem 1.3 implies that Snf
converges to the original function in Lp(·)([0,1))-norm (for more details see [8] and
the recent book [13], chapter 9).
In order to extend techniques and results of constant exponent case to the setting
of variable Lebesgue spaces, a central problem is to determine conditions on an
exponent p(·) under which the Hardy-Littlewood maximal operator is bounded on
Lp(·)(see monographs Cruz-Uribe and Fiorenza [2] and Diening et.al. [5]). We now
define the Hardy-Littlewood maximal function that is appropriate for the study of
Vilenkin-Fourier series. For fL1(G), let
Mf(x) = sup
xI,I∈F
1
µ(I)ZI
|f|dµ.
This maximal function was introduced first by P. Simon in [16]. He showed that
the maximal operator is bounded in Lp(G),1< p < and is of weak type (1,1).
Young [18] obtained the following analogue of Muckenhoupt’s theorem [11].
Theorem 1.4. Let wbe a weight function on G. For 1<p<, the following two
statements are equivalent:
(i) wAp(G),
摘要:

VILENKIN-FOURIERSERIESINVARIABLELEBESGUESPACESDAVITIADAMADZEANDTENGIZKOPALIANIAbstract.LetSnfdenotethenthpartialsumoftheVilenkin-Fourierseriesofafunctionf∈L1(G).For1

展开>> 收起<<
VILENKIN-FOURIER SERIES IN VARIABLE LEBESGUE SPACES DAVITI ADAMADZE AND TENGIZ KOPALIANI.pdf

共15页,预览3页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!
分类:图书资源 价格:10玖币 属性:15 页 大小:416.51KB 格式:PDF 时间:2025-05-06

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 15
客服
关注