On the mean value of certain functions connected with the convergence of orthogonal series


 Charlene McBride
 1 years ago
 Views:
Transcription
1 Analysis Mathematica, 4 (1978), On the mean value of certain functions connected with the convergence of orthogonal series B. S. KASIN Dedicated to Professor P. L. UVjanov on his 50th birthday In the study of the almost everywhere (a.e.) convergence of orthogonal series often arises the following question. Let Ф = Ф(и) = {<^(х)}" =1 be an orthonormal system of functions defined on the segment [0, 1]. Let us define the operator 1 S%: lg>~l 2 (Q, 1) in the following way. If у = {у^=1 б 1\, then (1) St(y) =f(x) = sup Let s(<p) denote the norm of the operator S%, i.e., (2) 5(Ф)= sup \\Si(y)h4o.i) llyllm^l We would like to estimate the number,у(ф). The classical result of D.E. Mensov and H. Rademacher (cf. [3, p. 188]) says that for every Ф = Ф(п) 2 we have (3) s(^)^clnw. For every n^l Mensov (cf. [3, p. 192]) exhibited an example of a system Ф 0 =Ф 0 (п) such that (4) 5(Ф 0 )^с1пл. In the present paper we shall study the "mean values" of the norms,у(ф). To illuminate the precise meaning of the term "mean value" we consider the following set of orthonormal systems 2" = {Ф}. Received March 1, By 11 we denote the space R n with Euclidean norm. 2 In the sequel С, с, and В denote positive absolute constants.
2 28 В. S. Kasin If Ф б", then Ф = {(р 1 (х)}" ==1 and each function <р ( (х) is piecewise constant on the n subintervals of [0, 1] of equal length. There is a onetoone correspondance between the systems <P Q n and the elements of the group O n of orthogonal matrices of order n. Namely, to the system Ф = {Ф^(Х)} corresponds the matrix А = {а^}^о п of the following form: (5),,.,==L^i^/i) il*l,j*n). With the aid of this correspondance the Haar measure fx n defined on O n (cf. [6, p. 43]) may be carried over onto the set Q n. In the sequel the measure of a closed set Ее Q n shall be denoted also by ц п (Е). We have the following Theorem. There exist absolute constants С and y>0 such that for every пш\ and t^o we have (6) M* fi": *(*) s* t} ^ (Се* 2 )". The statement of this theorem is included in [4] without proof. From this theorem we immediately obtain the corollaries below. Corollary 1. There exists an absolute constant В such that (б 7 )!л п {Феа п :8(Ф)^В}^е п. Corollary 2. Let S(n) denote the set of all permutations of the numbers 1,2,...,TI. For G S(n). let Ф^ denote the rearrangement of the terms of Ф ( 6") according to a. Then [хлф^о 1 : max $(Ф а ) шс Yhin} < n" n. From the estimation (6') it follows that there are "very few" systems having the property discovered by Mensov (we emphasize that in (4) the system Ф 0 (п) may be chosen from Q n ). For the proof of Theorem 1 we need some lemmas. First we shall introduce some notations. The number of elements of any finite set E shall be denoted by \E\. Furthermore, if x R m, then (x)j (l^j^m) denotes the jth coordinate of the vector x, and, as usually, R\ ={x = {(х)$ =1 : (x)j ^ 0, 1 ^ j ^ n}; S» = {x R n : \\х\\ п = 1}. If L is a subspace in R n, dim L = q 9 then m L denotes the normed Lebesgue measure on the gdimensional sphere S n DL (m L (S n f]l) = l). The measure m Rn (E) shall be denoted simply by m(e).
3 Convergence of orthogonal series 29 Lemma 1. For each rn^l one may choose a set of vectors Q m ={e} such that (i) GJ<C M ; (ii) if e Q m, then e R m + and e,sl/3; (Hi) for any vector y S m C)R"l there is a vector e=e(y) Q m such that (e(y))js(y)j (ls/sm). Proof. Suppose that y S m f}r1. Let us set for 5=0,1,2,... * E s (y) = {j : 2 1 < уj ё 2, 1 s ; = m}. It is easy to see that (7) a) b) ( i: г \E,(y)\ < 2* +», 2 /ST^j) 1/2 = ]^1 2 " For each j.s m n.r+ let us define the vector e(y) R m in the following way: (2* 1 if jсад, OSss [1 log 2 m] +1, (8) (e(y))j=\ (0 if ttl>e,(y), 0^ss[log 2 mj. From the construction of e{y) it is obvious that (9) 0^(e(y))j^(y)j. Besides, in virtue of the definition of E s (y) and part b) of relation (7) we have l f l J ' Let ««= U e(y). y S m nr On account of (9) and (9') to conclude the proof of Lemma 1 it is enough to show that O w ^C m. It is obvious that for any j;<es m ni^ and s^o we have \E s (y)\^m. Consequently, the number of the different systems Я (у) of the form 1 AGO = { 3,(y),..., GOD does not exceed (wf l) 2+(log2W)/2. From part a) of relation (7) it follows easily that the number of vectors e(y) of the system Q m with the same system k{y) does not exceed (O m П ст. l^$^(log 2 m)/2
4 30 В. S. Kasin Finally, for : the number \Q m \ we obtain the estimate l^s^(log 2w)/2 By using the elementary estimate C^^C q (m/q) q we obtain that a [(log 2 m)/2] ( m \ 22s [log 8 m] ( m \ ' sc " Ж Ш sc \g Ш 2q  This proves Lemma 1. { [log 2 m] / m \ 29 1 f[log 2 m] m 1 In Д (f j }= C m exp{ Д 2ЧпЩ g m\ 2 In [^C ffl exp{mc'}sc m. We shall use the following wellknown fact (cf, for example, [5, p. 335]). On the sphere S n there exists a set = {<} such that \Q' n \^C n and for any y S n there exists a vector e / = e / (y) Q / n for which (10) Ik'OOjlhsey. The estimate below is also wellknown. Lemma 2. For t^0 9 l^n<, and x 0 S n we have the inequality f(t) = m{x^s n : \(x,x 0 )\ <* t}^ e~^\ For t>l the statement of Lemma 2 is obvious, as in this case f(t)=0. For Q^t^l the set {x S n :(x, x 0 )^t} is contained in the hemisphere of radius (1 1 2 ) 1/2 and consequently, ДО (i./2)(ni)/2 = [(2 / 2 ) 1 /' 2 ]* аи (,,  1 >/ 2л ^ e~ f2n/4 (in obtaining the last inequality we used the estimation (1 xf^^e" 1, Lemma 2 has been proved. Lemma 2 immediately implies  0<x<:l). Consequence. Let X be a subspace of R\ d\mx=m, (Mx 0 i? w. Then for t^q we have (11) m x {* w fu: (x,x 0 )l ^/} ^ exp{ mr 2 W ^2}. Lemma 3. ITiere are absolute constants С and a>0 л/сй that for every subspace XaR n 0>3, diml^ 1+«/2), vector a={a f } *S r ", and number t^o we have the inequality (12) m L fj? 5"nL: max l^e, J ^ Л ^ Cfe" 1 ".
5 Convergence of orthogonal series 31 Proof. It is enough to prove Lemma 3 for t^20jn w, since for t^20n~ 1/2 the statement of our lemma is obtained because of the choice of the constant С in (12). Without loss of generality we may suppose that a^o (1^Шп). n For $=0,1,2,...' let us split the sum 2ад i nt0 the "pieces" {P*}^1 in the following way: n Hi 1 n Р =2 1Уи Pl= 2 ЪУг + Яп^т* Р1 = (а П1 ЯпОУт+ 2 а гугi=l i=l i=n 1 +l Analogously, the "piece" P* 1 is represented in the form ps 1 ps _j_ ps r v  r.2v't"' r 2v + l' A similar division is used in the proof of Erdos' theorem on the a.e. convergence of lacunary trigonometric series (cf. [1, p ]). It is not hard to see (cf. [1, p. 705, Lemma 2]) that the splitting {PUfJi 1, $=0, 1, 2,..., of the above form may be chosen in such a way that for each "piece" n P s v = 2 a \ v,s) yt (0^v<2") we have the estimation i=l (13) J[a/ V ' s) ] 2^2" s. 8 = 1. (13) implies that we can find an integer s 0 such that all "pieces" P* (0^v^2* 1) contain at most two terms. Then we have the inequality max 2^Уг so ^ 2 max \P S V\ + max а^,. 0^v<2 s=0 = v ' ля s 1^Шп Consequently, if for fixed vectors {a} and {y} and for s=0,l,... 9 s 0 inequalities (14) then Hence we have max \P S V 0^V<2 5 max VII r 2 i = l and max \a t y t \ ^, 5(s+l) 2 iiiiv ' "'  3 we have the (15) m L \ytst\l: max 24j#' = 2 2 may^s'^l.^^jt ^]^ + т ь {^ 5"П :шах в у, *4}.
6 .32 В, S. Kasin Due to (13), the consequence of Lemma 2 (cf. (11)), and the inequality diml^ ^ l+n/2 we obtain that <16> У^{у ЬП8':\Р'\^1^}^^{ 1^^У Besides, by using (11) we get that {16') m L \yzst\l: max \a,y t \ ё 4"} = Si Wt {,^nx: W>^} S iexp{0}. Comparing inequalities (15), (16), and (16') we arrive at <17) mjye^nl: max 24yJ = Л = io Г 2 s / 2 ra 1 " ( t 2 n] If г 2 и>200, then it is easy to show that s o Г 2 s / 2 n 1 and, moreover, ^ех Н~^Л~^< ^ ехр ^4 j?2«exp{~2^с'еу* п (у > 0). By putting the last inequalities into (17) we obtain the estimation (12). Lemma 3 is proved. We are going to use the following property of the measure /л п on the group O n (cf. [6, p. 55]). Let usfixan arbitrary vector e 0 S n and denote by О й ~ 1 = {Л) the subgroup of the group O n whose elements are those matrices A for which Ae 0 =e 0. The symbol fi n^l9 as before, denotes Haar measure on О п ~ г. For e S" let us choose an arbitrary matrix A e from O n with A e e 0 = e. Then we have {18) ff(a) dfi n = J dm f f(a e A) dii n. x $n 0nl for any measurable bounded function /04), A O n.
7 Convergence of orthogonal series 33 Proof of Theorem. We may suppose that л^10, as for ж 10 the statement of the theorem shall be satisfied if we choose the constant С in (6) large enough. Let a system Ф () п be given. Suppose that the matrix А(Ф)=А = {а (] } О п is defined by the equality (5). Then it is easy to see that 2 \2у* а Л\ = sup 2 2 y + 2 \2УЫ\ # y={ yi }<=S*,{Nj}4 = l \i = l t i=[!t/2] + l Vi = l /J Consequently, it is enough to prove that for m=[n/2] and m=n [n/2] and for all t^o we have 2л 1/2 (20) M(t) = ti{aco»: '^ sup [J (i y ; a y j 2 ] 1/2 is L} ^ (Cei**)" Let the sequence {Д/}" =1 of numbers be fixed. By using the properties of the system of vectors Q' n (cf. (10)) we obtain that [ m / AT, \2i"ii/a Г m f N J /2 2 2w s sup 2\2(уе'(у)Ш,^_ j=l \i = l /.J ye^lj^l 4=1 ^2, 1/2 + supfi{f(e'0')) J a y } 8 ] 3,^S n Lj=i lj=i J J f и ч1/2 Taking into account that Z(^~^(7)) 2 =1/2 we arrive at 1 г ш г NJ i 2 ii/ 2 J J As ^ ^С И (cf. (10)), from the last inequality and from the definition of the function M{t) (cf. (20)) we obtain that (21) M(t) * {СУ sup JASO*: sup f 2 [2 *,<J1 =" 4=}. By virtue of Lemma 1 from inequality (21) it follows that (22) M(t) ^C n sup /*jv< ц \ A O n : sup zes n,e Q m I {#,} J? 2ъ<*! j i=l j=(e)j,l*j*m Let the vectors z^s", e Q m, and the number t^0 befixed.then (23) /iu 0":sup 1 3 Analysis Mathematica t 1 =(e) y, lijsm = / {ty} V2 * nn o n J /=i = m } nx )dn,
8 34 В. S. Kasin where A = {a ij }eo n, l^j^rn, and I 1 if sup {0 otherwise. N j 2 z^ij i=l Jf^> As \\e\\i^l/3 9 е О т (cf. Lemma 1), a multiple application of Lemma 3 and equality (18) show that f л m m ( а л (24) f IIxM)dfi n = ncap\ T t*n(ejn* on j=i j=i { z J ^ C m Qxpljt 2 n 2(e)j\ ^ C n exp(^2) 5 where y>0 is an absolute constant. From (22) (24) we obtain that M(t)^ Cexp{ynt 2 }. Thus inequality (20) is proved. As it was noted earlier, the statement of our theorem follows from (20). Remark. We estimated the mean value of the function я(ф) on the set Q n consisting of orthonormal systems. The estimation of the mean value of the function,у(ф) on the set of all (not necessarily orthogonal) systems such that Ф {(Pi(x)}" =lr> \<p.(x)\ = l 9 the function (р г (х) being constant on the intervals ((j l)/n, j/n), t^i, j=n, x [0, 1], may be carried out following the method of [2]. References [1] H. К. Бари, Тригонометрические ряды, Физматгиз (Москва, 1961). [2] G. BENNETT, V. GOODMAN, С. М. NEWMAN, Norms of random matrices, Pacific J. Math., 59 (1975), [3] S. KACZMARZ und H. STEINHAUS, Theorie der Orthogonalreihen (Warszawa Lwow, 1935) С. Качмаж и Г. Штейнгауз, Теория ортогональных рядов, Физматгиз (Москва, 1958). [4] Б. С. Кашин, О некоторых свойствах ортогональных систем сходимости, Труды матем. инта им. В. А. Стеклова АН СССР, 143 (1977), [5] L. FEJES TOTH, Lagerungen in der Ebene, aufder Kugel und im Raum, Springer (Berlin Gottingen Heidelberg, 1953) Л. Фейеш Т от, Расположения на плоскости, на сфере и в пространстве, Физматгиз (Москва, 1958). [6] A. WEIL, LHntegration dans les groupes topologiques et ses applications (Paris, 1940) А. Вей ль, Интегрирование в топологических группах и его применения, Иностранная литература (Москва, 1950).
9 Convergence of orthogonal series 35 О среднем значении некоторых функций, связанных со сходимостью ортогональных рядов Б. С. КАШИН В статье рассматриваются множества Q n, 1 ^ п < «*>, ортонормйрованных систем Ф = = i<pi ( *)}"= 1» состоящих из фушспжйэ постоянных на интервалах I» I» 1 ^j^n. На 6 й естественно переносится с группы ортогональных матриц порядка п мера Хаара. Изучается поведение # на Q n функции 8(Ф)= sup (/ sup (iyiviidfdxf*. Доказывается, что при / > 0 и п = 1, 2,... /*{Ф Й П : л(ф) ^ /} ^ (Се у **) я. Б. С. КАШИН СССР, МОСКВА УЛ. ВАВИЛОВА 42 МАТЕМАТИЧЕСКИЙ ИНСТИТУТ ИМ. В. А. СТЕКЛОВА АН СССР 3*
An example of a computable
An example of a computable absolutely normal number Verónica Becher Santiago Figueira Abstract The first example of an absolutely normal number was given by Sierpinski in 96, twenty years before the concept
More informationINTEGRAL OPERATORS ON THE PRODUCT OF C(K) SPACES
INTEGRAL OPERATORS ON THE PRODUCT OF C(K) SPACES FERNANDO BOMBAL AND IGNACIO VILLANUEVA Abstract. We study and characterize the integral multilinear operators on a product of C(K) spaces in terms of the
More informationChapter 5. Banach Spaces
9 Chapter 5 Banach Spaces Many linear equations may be formulated in terms of a suitable linear operator acting on a Banach space. In this chapter, we study Banach spaces and linear operators acting on
More informationTHE KADISONSINGER PROBLEM IN MATHEMATICS AND ENGINEERING: A DETAILED ACCOUNT
THE ADISONSINGER PROBLEM IN MATHEMATICS AND ENGINEERING: A DETAILED ACCOUNT PETER G. CASAZZA, MATTHEW FICUS, JANET C. TREMAIN, ERIC WEBER Abstract. We will show that the famous, intractible 1959 adisonsinger
More information1 Sets and Set Notation.
LINEAR ALGEBRA MATH 27.6 SPRING 23 (COHEN) LECTURE NOTES Sets and Set Notation. Definition (Naive Definition of a Set). A set is any collection of objects, called the elements of that set. We will most
More informationInner Product Spaces and Orthogonality
Inner Product Spaces and Orthogonality week 34 Fall 2006 Dot product of R n The inner product or dot product of R n is a function, defined by u, v a b + a 2 b 2 + + a n b n for u a, a 2,, a n T, v b,
More informationFixed Point Theorems and Applications
Fixed Point Theorems and Applications VITTORINO PATA Dipartimento di Matematica F. Brioschi Politecnico di Milano vittorino.pata@polimi.it Contents Preface 2 Notation 3 1. FIXED POINT THEOREMS 5 The Banach
More informationA UNIQUENESS RESULT FOR THE CONTINUITY EQUATION IN TWO DIMENSIONS. Dedicated to Constantine Dafermos on the occasion of his 70 th birthday
A UNIQUENESS RESULT FOR THE CONTINUITY EQUATION IN TWO DIMENSIONS GIOVANNI ALBERTI, STEFANO BIANCHINI, AND GIANLUCA CRIPPA Dedicated to Constantine Dafermos on the occasion of his 7 th birthday Abstract.
More informationSET THEORY AND UNIQUENESS FOR TRIGONOMETRIC SERIES. Alexander S. Kechris* Department of Mathematics Caltech Pasadena, CA 91125
SET THEORY AND UNIQUENESS FOR TRIGONOMETRIC SERIES Alexander S. Kechris* Department of Mathematics Caltech Pasadena, CA 91125 Dedicated to the memory of my friend and colleague Stelios Pichorides Problems
More informationON A PBOBLEM OF FOKMAL LOGIC
264 F. P. KAMSEY [Dec. 13, ON A PBOBLEM OF FOKMAL LOGIC By F. P. KAMSEY. [Received 28 November, 1928. Read 13 December, 1928.] This paper is primarily concerned with a special case of one of the leading
More informationU.C. Berkeley CS276: Cryptography Handout 0.1 Luca Trevisan January, 2009. Notes on Algebra
U.C. Berkeley CS276: Cryptography Handout 0.1 Luca Trevisan January, 2009 Notes on Algebra These notes contain as little theory as possible, and most results are stated without proof. Any introductory
More information3. INNER PRODUCT SPACES
. INNER PRODUCT SPACES.. Definition So far we have studied abstract vector spaces. These are a generalisation of the geometric spaces R and R. But these have more structure than just that of a vector space.
More informationANALYTIC NUMBER THEORY LECTURE NOTES BASED ON DAVENPORT S BOOK
ANALYTIC NUMBER THEORY LECTURE NOTES BASED ON DAVENPORT S BOOK ANDREAS STRÖMBERGSSON These lecture notes follow to a large extent Davenport s book [5], but with things reordered and often expanded. The
More informationBasic Concepts of Point Set Topology Notes for OU course Math 4853 Spring 2011
Basic Concepts of Point Set Topology Notes for OU course Math 4853 Spring 2011 A. Miller 1. Introduction. The definitions of metric space and topological space were developed in the early 1900 s, largely
More informationMA651 Topology. Lecture 6. Separation Axioms.
MA651 Topology. Lecture 6. Separation Axioms. This text is based on the following books: Fundamental concepts of topology by Peter O Neil Elements of Mathematics: General Topology by Nicolas Bourbaki Counterexamples
More informationA Characterization of Affine Surface Area
A Characterization of Affine Surface Area Monika Ludwig and Matthias Reitzner Abteilung für Analysis, Technische Universität Wien, Wiedner Hauptstraße 810/114, A1040 Vienna m.ludwig+e114/m.reitzner+e114@tuwien.ac.at
More informationReading material on the limit set of a Fuchsian group
Reading material on the limit set of a Fuchsian group Recommended texts Many books on hyperbolic geometry and Kleinian and Fuchsian groups contain material about limit sets. The presentation given here
More informationGroup Theory. Contents
Group Theory Contents Chapter 1: Review... 2 Chapter 2: Permutation Groups and Group Actions... 3 Orbits and Transitivity... 6 Specific Actions The Right regular and coset actions... 8 The Conjugation
More informationLEARNING OBJECTIVES FOR THIS CHAPTER
CHAPTER 6 Woman teaching geometry, from a fourteenthcentury edition of Euclid s geometry book. Inner Product Spaces In making the definition of a vector space, we generalized the linear structure (addition
More informationOn an antiramsey type result
On an antiramsey type result Noga Alon, Hanno Lefmann and Vojtĕch Rödl Abstract We consider antiramsey type results. For a given coloring of the kelement subsets of an nelement set X, where two kelement
More informationA new continuous dependence result for impulsive retarded functional differential equations
CADERNOS DE MATEMÁTICA 11, 37 47 May (2010) ARTIGO NÚMERO SMA#324 A new continuous dependence result for impulsive retarded functional differential equations M. Federson * Instituto de Ciências Matemáticas
More informationFixed Point Theorems For SetValued Maps
Fixed Point Theorems For SetValued Maps Bachelor s thesis in functional analysis Institute for Analysis and Scientific Computing Vienna University of Technology Andreas Widder, July 2009. Preface In this
More informationPROOFS BY DESCENT KEITH CONRAD
PROOFS BY DESCENT KEITH CONRAD As ordinary methods, such as are found in the books, are inadequate to proving such difficult propositions, I discovered at last a most singular method... that I called the
More informationRegret in the OnLine Decision Problem
Games and Economic Behavior 29, 7 35 (1999) Article ID game.1999.0740, available online at http://www.idealibrary.com on Regret in the OnLine Decision Problem Dean P. Foster Department of Statistics,
More informationCONTRIBUTIONS TO ZERO SUM PROBLEMS
CONTRIBUTIONS TO ZERO SUM PROBLEMS S. D. ADHIKARI, Y. G. CHEN, J. B. FRIEDLANDER, S. V. KONYAGIN AND F. PAPPALARDI Abstract. A prototype of zero sum theorems, the well known theorem of Erdős, Ginzburg
More informationTAKEAWAY GAMES. ALLEN J. SCHWENK California Institute of Technology, Pasadena, California INTRODUCTION
TAKEAWAY GAMES ALLEN J. SCHWENK California Institute of Technology, Pasadena, California L INTRODUCTION Several games of Tf takeaway?f have become popular. The purpose of this paper is to determine the
More informationFixed Point Theorems in Topology and Geometry
Fixed Point Theorems in Topology and Geometry A Senior Thesis Submitted to the Department of Mathematics In Partial Fulfillment of the Requirements for the Departmental Honors Baccalaureate By Morgan Schreffler
More informationNOTES ON GROUP THEORY
NOTES ON GROUP THEORY Abstract. These are the notes prepared for the course MTH 751 to be offered to the PhD students at IIT Kanpur. Contents 1. Binary Structure 2 2. Group Structure 5 3. Group Actions
More informationIRREDUCIBLE OPERATOR SEMIGROUPS SUCH THAT AB AND BA ARE PROPORTIONAL. 1. Introduction
IRREDUCIBLE OPERATOR SEMIGROUPS SUCH THAT AB AND BA ARE PROPORTIONAL R. DRNOVŠEK, T. KOŠIR Dedicated to Prof. Heydar Radjavi on the occasion of his seventieth birthday. Abstract. Let S be an irreducible
More informationThe whole of analytic number theory rests on one marvellous formula due to Leonhard Euler (17071783): n s = primes p. 1 p
Chapter Euler s Product Formula. The Product Formula The whole of analytic number theory rests on one marvellous formula due to Leonhard Euler (707783): n N, n>0 n s = primes p ( p s ). Informally, we
More information