ON THE DISTRIBUTION OF k-th POWER FREE INTEGERS, II

Similar documents
Research Article A Limit Theorem for Random Products of Trimmed Sums of i.i.d. Random Variables

Analytic Number Theory Solutions

A REFINEMENT OF JENSEN S INEQUALITY WITH APPLICATIONS. S. S. Dragomir 1. INTRODUCTION

Chapter 2 Transformations and Expectations

Solutions to Problem Sheet 1

Songklanakarin Journal of Science and Technology SJST R1 Teerapabolarn

Negative values of truncations to L(1, χ)

PRIME RECIPROCALS AND PRIMES IN ARITHMETIC PROGRESSION

1 = 2 d x. n x n (mod d) d n

A Simple Proof Of The Prime Number Theorem N. A. Carella

ECE534, Spring 2018: Final Exam

Notes on the prime number theorem

Proposition 2.1. There are an infinite number of primes of the form p = 4n 1. Proof. Suppose there are only a finite number of such primes, say

Matrix Operators and Functions Thereof

Chapter 2. Finite Fields (Chapter 3 in the text)

Ma 4121: Introduction to Lebesgue Integration Solutions to Homework Assignment 5

k=1 s k (x) (3) and that the corresponding infinite series may also converge; moreover, if it converges, then it defines a function S through its sum

Lecture 6 Testing Nonlinear Restrictions 1. The previous lectures prepare us for the tests of nonlinear restrictions of the form:

arxiv: v4 [math.co] 5 May 2011

6.451 Principles of Digital Communication II Wednesday, March 9, 2005 MIT, Spring 2005 Handout #12. Problem Set 5 Solutions

Logarithm of the Kernel Function. 1 Introduction and Preliminary Results

Convergence of random variables. (telegram style notes) P.J.C. Spreij

arxiv: v1 [math.nt] 5 Jan 2017 IBRAHIM M. ALABDULMOHSIN

Sketch of Dirichlet s Theorem on Arithmetic Progressions

Ma 530 Infinite Series I

Representing Functions as Power Series. 3 n ...

( a) ( ) 1 ( ) 2 ( ) ( ) 3 3 ( ) =!

A Central Limit Theorem for Belief Functions

Equations and Inequalities Involving v p (n!)

It is often useful to approximate complicated functions using simpler ones. We consider the task of approximating a function by a polynomial.

3.1. Introduction Assumptions.

A Note on Sums of Independent Random Variables

PERIODS OF FIBONACCI SEQUENCES MODULO m. 1. Preliminaries Definition 1. A generalized Fibonacci sequence is an infinite complex sequence (g n ) n Z

The Asymptotic Expansions of Certain Sums Involving Inverse of Binomial Coefficient 1

Abelian Property of the Category of U-Complexes

ON SOME NEW SEQUENCE SPACES OF NON-ABSOLUTE TYPE RELATED TO THE SPACES l p AND l I. M. Mursaleen and Abdullah K. Noman

Weak and Strong Convergence Theorems of New Iterations with Errors for Nonexpansive Nonself-Mappings

Complex Analysis Spring 2001 Homework I Solution

Analytic Continuation

THE INTEGRAL TEST AND ESTIMATES OF SUMS

On the asymptotic growth rate of the sum of p-adic-valued independent identically distributed random variables

Gaps between Consecutive Perfect Powers

A Note on Bilharz s Example Regarding Nonexistence of Natural Density

VIETA-LIKE PRODUCTS OF NESTED RADICALS

Title. Author(s)Cho, Yonggeun; Jin, Bum Ja. CitationJournal of Mathematical Analysis and Applications, 3. Issue Date Doc URL.

THE GREATEST ORDER OF THE DIVISOR FUNCTION WITH INCREASING DIMENSION

DANIELL AND RIEMANN INTEGRABILITY

Sparsification using Regular and Weighted. Graphs

The structure of Fourier series

Almost all hyperharmonic numbers are not integers

ALGEBRA HW 7 CLAY SHONKWILER

Definition 2 (Eigenvalue Expansion). We say a d-regular graph is a λ eigenvalue expander if

1. C only. 3. none of them. 4. B only. 5. B and C. 6. all of them. 7. A and C. 8. A and B correct

Two Topics in Number Theory: Sum of Divisors of the Factorial and a Formula for Primes

COMPUTING FOURIER SERIES

The natural exponential function

sin(n) + 2 cos(2n) n 3/2 3 sin(n) 2cos(2n) n 3/2 a n =

A Note on the Weak Law of Large Numbers for Free Random Variables

2.4.2 A Theorem About Absolutely Convergent Series

On Elementary Methods to Evaluate Values of the Riemann Zeta Function and another Closely Related Infinite Series at Natural Numbers

Final Solutions. 1. (25pts) Define the following terms. Be as precise as you can.

MATH 6101 Fall 2008 Series and a Famous Unsolved Problem

On triangular billiards

Exponential function and its derivative revisited

An analog of the arithmetic triangle obtained by replacing the products by the least common multiples

Non-Archimedian Fields. Topological Properties of Z p, Q p (p-adics Numbers)

Dirichlet s Theorem on Arithmetic Progressions

Math 210A Homework 1

ECE 330:541, Stochastic Signals and Systems Lecture Notes on Limit Theorems from Probability Fall 2002

Lommel Polynomials. Dr. Klaus Braun taken from [GWa] source. , defined by ( [GWa] 9-6)

Sequences, Mathematical Induction, and Recursion. CSE 2353 Discrete Computational Structures Spring 2018

Sequences A sequence of numbers is a function whose domain is the positive integers. We can see that the sequence

Harmonic Number Identities Via Euler s Transform

Q-BINOMIALS AND THE GREATEST COMMON DIVISOR. Keith R. Slavin 8474 SW Chevy Place, Beaverton, Oregon 97008, USA.

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4.

Assignment 5: Solutions

ENGI Series Page 6-01

Physics 116A Solutions to Homework Set #1 Winter Boas, problem Use equation 1.8 to find a fraction describing

Integrable Functions. { f n } is called a determining sequence for f. If f is integrable with respect to, then f d does exist as a finite real number

Mathematical Methods for Physics and Engineering

MATH 6101 Fall Problems. Problems 11/9/2008. Series and a Famous Unsolved Problem (2-1)(2 + 1) ( 4) 12-Nov-2008 MATH

Fixed Points Theorems In Three Metric Spaces

Approximation properties of (p, q)-bernstein type operators

Comparison Study of Series Approximation. and Convergence between Chebyshev. and Legendre Series

6.3 Testing Series With Positive Terms

APPROXIMATION OF CONTIONUOUS FUNCTIONS BY VALLEE-POUSSIN S SUMS

The Riemann Zeta Function

Extremality and Comparison Results for Discontinuous Third Order Functional Initial-Boundary Value Problems

Weighted Gcd-Sum Functions

Confidence Intervals

Weil Conjecture I. Yichao Tian. Morningside Center of Mathematics, AMSS, CAS

Modern Algebra 1 Section 1 Assignment 1. Solution: We have to show that if you knock down any one domino, then it knocks down the one behind it.

Archimedes - numbers for counting, otherwise lengths, areas, etc. Kepler - geometry for planetary motion

YALE UNIVERSITY DEPARTMENT OF COMPUTER SCIENCE

k-equitable mean labeling

Solutions to Problem Set 7

An elementary proof that almost all real numbers are normal

A 2nTH ORDER LINEAR DIFFERENCE EQUATION

ABSOLUTE CONVERGENCE OF THE DOUBLE SERIES OF FOURIER HAAR COEFFICIENTS

Transcription:

Duy, T.K. a Taaobu, S. Osaa J. Math. 5 (3), 687 73 O THE DISTRIBUTIO OF -TH POWER FREE ITEGERS, II TRIH KHAH DUY a SATOSHI TAKAOBU (Receive July 5,, revise December, ) Abstract The iicator fuctio of the set of -th ower free itegers is aturally etee to a raom variable X () ( ) o ( Ç, ), where Ç is the rig of fiite itegral aeles a is the Haar robability measure. I the revious aer, the first author ote the strog law of large umbers for {X () ( )}, a showe the asymtotics: E [(Y () ) ] as, where Y () () Ï È X () ( ) (). I the reset aer, we rove the covergece of E [(Y () ) ]. For this, we reset a geeral roositio of aalytic umber theory, a give a roof to this.. Itrouctio Let Ç be the rig of fiite itegral aeles; B the Borel -fiel of Ç ; the Haar robability measure o (Ç,B). I [4, ], the trilet ( Ç,B, ) is itrouce i the followig way: For a rime umber, the -aic metric o is efie by (, y) Ï if{ l Á l ( y)},, y ¾. The comletio of by is eote by. By eteig the algebraic oeratios a i cotiuously to those i, the comact metric sace (, ) becomes a rig. I articular, (, ) is a comact abelia grou with resect to. Thus, there is a uique Haar robability measure with resect to o (,B( )), where B( ) is the Borel -fiel of. Puttig i i-th rime umber (i,, ), we set Ç Ï Ï i For ( i ), y (y i ) ¾ Ç, we efie i, (, y) Ï i i i i ( i, y i ), i. y Ï ( i y i ), y Ï ( i y i ). Mathematics Subject Classificatio. Primary 6F5; Secoary 6B, 6B5, 37, K4.

688 T.K. DUY AD S. TAKAOBU By these efiitios, Ç becomes a rig, which is just the rig of fiite itegral aeles state above. (Ç, ) is agai a comact metric sace, a both a are cotiuous. I articular, this is a comact abelia grou with resect to, a its Haar robability measure is othig but. By ietifyig with the iagoal set {(,, ) ¾ Á ¾ } Ç, it is see that is a ese subrig of Ç. Thus Ç is a comactificatio of. Let be a iteger,. Let B () be the set of all elemets i Ç havig o -th ower factors, i.e., B () Ï { ¾ Ç Á (: rime)}, where ¾ Ç (, so ¾ Ç ÒÇ ), a X () Ï B () fuctio of B () ). The followig are results of Duy []: ( the iicator È Fact (Strog law of large umbers). lim () X () () (), -a.e.. Here ( ) is the Riema zeta fuctio. () For each ¾ Æ, we set Y () () Ï () X () ( ) (). Fact. E [(Y () ) ] as. Fact 3. A sequece {Y () } (Ç i L,B, ) has o limit oit. amely, for ay subsequece { i } i (), {Y i } i is ot coverget i L as i. Fact follows at oce from the ergoicity of the shift È a E [X () ] (). From this fact, we have the followig questio: Whe (X () ( ) ()) is ormalize aroriately, is its istributio wealy coverget as? ¾ Fact tells us that a ormalizig costat must be (), a that a sequece {(Y () of istributios o Ê is tight. Fact 3 is a fuctioal aalytical result a brigs )} o ews for the behavior of Y () as. But, for this, we eect to have a limit theorem i robability theory. (Ufortuately, we still have o iformatio o this limit theorem.) I this aer, we mae some remar about Fact a Fact 3. Cf. Æ i the roof of Claim.

Theorem. lim E [(Y () ) ] -TH POWER FREE ITEGERS II 689 ( ) () () si(()). Theorem. (i) lim lim M E [(Y () M Y () ) ] É ( )( ) ( )(() ()si()). Fact 3 above is a cosequece of this. (ii) But, a whole sequece {Y () } (Ç i L,B,) is wealy coverget to as. Throughout È É this aer, the letter eotes a rime umber, a the symbols a are a rouct a a summatio etee over all rime umbers, resectively. Theorems above will be rove i Sectio 4. I Sectio, a aother comutatio of E [Y () M Y () ], which is ifferet from oe i Duy [], is give. A, i Sectio 3, to rove Theorem, we reare Proositio. This is a geeral roositio of aalytic umber theory, a will be rove i Sectio 5. The authors woul lie to tha the referee for goo avice which eable us to mae roofs clear a cosierably short.. Comutatio of E [Y () M Y() ] By a ifferet aroach from Duy [], we comute E [Y () M Y () ]. Claim. For M, E [Y () M Y () ] M () () c (c) c M c c M c c. Here ( ) is the Möbius fuctio a {a} is the fractioal art of the real umber a. Proof. Æ First E [Y () M Y () ] E M () () Fi M. We ivie the roof ito three stes: M m X () ( m) () X () ( ) () Duy s metho is origially ue to [4]. The same i of comutatio i the roof of Claim aears i early stuy of [4]. So, a hrase ifferet aroach may be too much to say.

69 T.K. DUY AD S. TAKAOBU M () () otig that mm, E [X () ( m)x () ( )] () (E [X () ( m)] E [X () ( )]) (). () X () (y) ( (y)), where, for ¾ Æ, (y) Ï, y ( y ¾ Ç ),, otherwise, (3) (4) we have { } is ieeet, ( ), ( ), E [X () ( m)] E [X () ()] (by the shift ivariace of ) () (by Euler s rouct of ( )), a thus Sice, by () E [Y () M Y () ] M () () mm, E [X () ( m)x () ( )] X () ( m)x () ( ) ( ( m)) ( ( )) (). ( ( ) ( m) ( m) ( )) ( ( ) ( m) (m ) ( )) (by a ietity: ( m) ( ) (m ) ( )),

we see from (3) a (4) that -TH POWER FREE ITEGERS II 69 E [Y () M Y () ] M () () mm, (m ) (). Æ By Euler s rouct of ( ) () () () c ( ) (c) c (where () Ï #{Á } the umber of ifferet rime factors of ) () c c, (c ) ( ) (c) c (c ) ( ) ( ) c (there eists a oe-to-oe corresoece betwee the set {(c, )Á is square free a c } a the set {(c, )Á c is square free}; a corresoece from the former to the latter is (c, ) (c, c) a oe from the latter to the former is (c, ) (c, c ). Here (c, ) a (c, ) eote a air of c a, a that of c a, resectively) (c ) ( ) ( ) c (c, ) c, (where (c, ) is the greatest commo ivisor of c a. ote that (c ) if (c, ) ) (c )( ) ( ) ( ) c (c, ) c,

69 T.K. DUY AD S. TAKAOBU (by the multilicativity of ) (c ) c c (c ) c c ( ) (c, )( ) ( ) (c, )( ) () () (by the multilicativity of ( ) (c, )( ) ( ) ) c (c ) c. c Similarly, sice (m ) (c ) c c c (m ), c we have c (m ) (c) c c () c (m ) c. Thus, by Æ a 3 Æ below E [Y () M Y () ] M () () mm, (c) M () () c c (c) M () () c c This is the assertio of the claim. c (c) c c (m ) c c M c m Mc c c (m ) c M c c.

-TH POWER FREE ITEGERS II 693 3 Æ Fi u ¾ Æ. Let Q a s be a quotiet a a remaier of ivie by u, resectively. Thus Qu s, where Q u 3, s { u}u ¾ {,,, u }. The Q u (m ) q j Q q j Q u u (m ((q )u j)) s u u q j Q u j Q s j u (m (Qu j)) u (m j (q )u) s u (m j Qu) j u (m j) s u (m j) j u (m j) s u (m j) j u (m j) j (by a ietity: u (y u) u (y)) È (first u j u(m j) È ju u(m j) È ju u(m mo u j), where m mo u Ï the remaier of m ivie by u. Secoly, otig that for j u, u (m mo u j) j m mo u (mo u) j È u m (mo u), we see j u(m j)) u Therefore u M m u u s M m j u (m j). u (m ) u M m u u (m ) M u s j u (m j) M 3 For a ¾ Ê, a Ï ma{ ¾ Á a} a a Ï mi{ ¾ Á a }. We call Ï Ê a Ï Ê the floor fuctio a the ceilig fuctio, resectively. ote that {a} a a ¾ [, ). u

694 T.K. DUY AD S. TAKAOBU u u u u M u s M j m M u s M u r j i M u M s u r s i j M u M s u r s u ( j m) M u u ( j i) M u u (i j) M M u u (where r {Mu}u) (for i, j u, u i j u. Also u (i j) i j (mo u). Thus u (i j) i j) r s M u u u r s r s u u u u M u Claim. u u M u u s u M u (because {Mu} r u, { u} su) M u u For each ¾ Æ, lim M E [Y () M Y () ]. Proof. Let M. Sice {Mc }, { c }, M c M c c c c c. (by a ietity: ab (a b)(a b)). Multilyig both sies by (M () )( () )(c) É c ( ), a the aig them over c ¾ Æ yiel that E [Y () M Y () ] M () () M c (c) () E [(Y () ) ]. c From this, the assertio of the claim follows. c c

3. Presetatio of Proositio -TH POWER FREE ITEGERS II 695 (5) By Claim E [(Y () ) ] c c É (c) c ( ) f (c) c c c c, where (6) f () Ï () É ( ), ¾ Æ. To show the covergece of E [(Y () ) ] as a to fi the value of this limit, we reset a geeral roositio: Proositio. coitio (7) or (8): (7) (8) Let a arithmetic fuctio f, i.e., fï Æ satisfy the followig () f, su ¾Æ f (), È f has the mea-value M( f ), i.e., lim () f () is coverget to a fiite limit M( f ). The, it hols that for ¾ (, ) 4 a h ¾ C [, ] with h() (9) lim f ()h M( f ) h({}). Before rovig this roositio, we give some commets o the coitios (7) a (8): Claim 3. If f Ï Æ satisfies the coitio (7), the f has the mea-value M( f ) () f. 4 Here may be a real umber,, though was a iteger, at the begiig of this aer.

696 T.K. DUY AD S. TAKAOBU Proof. For simlicity, we efie f Ï Æ by () f () () f, ¾ Æ. Sice, by the Möbius iversio formula () f () f (), we have for, y ¾ [, ) f () Á f () f () f () y f () { } f () f () { } f (). y { } f () Trasosig the first term of the last right-ha sie, a the taig the absolute value, we see that () f () f () y f () y f () f () y y y { f () y } f () f (). y { } f () f () y By lettig a y, the assertio of the claim follows. REMARK. Schwarz Siler [3] calls Claim 3 Witer s theorem.

-TH POWER FREE ITEGERS II 697 We give a eamle of f satisfyig the coitio (7): Let fï Æ be multilicative, i.e., f a f (m) f (m) f () EXAMPLE. rovie that (m, ). If, i aitio, (3) f () the f satisfies the coitio (7)., l f ( l ) l, Proof. Multilicativity of a f is iherite to f, a so f. I geeral, multilicativity of a arithmetic fuctio imlies a rouct reresetatio of Dirichlet series associate with the fuctio. Thus Sice, by () (4) (5) f () e f () f ( ) f () µ f ( ) l l f () () f ()() f () f f () (ote that f () ), l (by a iequality: e ( ¾ Ê)). f ( l ) () f ( l )() f ( l ) (ote that ( j ) ( j )) f ( l ) f ( l ) (l ), we have f () f () f () l f ( l ) l l f () l f ( l ) f ( l ) l f ( l ) l l l f ( ) l l f () f ( l ) l

698 T.K. DUY AD S. TAKAOBU 3 f () (by (3)). Therefore f satisfies the coitio (7). l f ( l ) l The coitio (7) oes ot always imly the coitio (8). EXAMPLE. Let f Ï Æ be multilicative, a satisfy f () «, f (l ) (l ) for each rime, where «¾ (, ). Sice f (), «f ( ) satisfies the coitio (7) from Eamle. Also, sice f ( m ) f ( ) f ( m ) we see that This imlies that m i «i m m i i e «i e È m i «i, e («i )(«i ) (by a iequality: log( ) () ( )) È m e i («i )(«i È ) e («( «m )) i «i, lim f ( m if «, ) m if «. su f () if «, if «.

4. Proof of two theorems Proof of Theorem. satisfies Also, ote that -TH POWER FREE ITEGERS II 699 f Ï Æ, efie by (6), is clearly multilicative, a f () l f ( ) l l. f () e È 4 ( ¾ Æ), (because, by ( )4, É ( ) É (4 ) É e4 e È 4 e È 4 ). Hece, this f ( ) satisfies both the coitio (7) a the coitio (8), so that alyig Proositio, we see (6) lim c f (c) c c M( f ) {}( {}) Let f be a multilicative fuctio efie by (). By (4) a (5) f (), f ( l ), l,, l 3 for rime a iteger l,. Claim 3 the imlies that M( f ) f () f () l f ( l ) l..

7 T.K. DUY AD S. TAKAOBU Collectig (5), (6) a this, we have lim E [(Y () ) ] {}( {}). Let us fi the value of a itegral o the right-ha sie. The Fourier easio of a fuctio {}( {}) is as follows: {}( {}) 6 cos cos si. (because È 6) Termwise itegratio yiels that {}( {}) si si y y (y) si y y y si y y y We here ote that from a formula: Ê (si Ú) u Ú u ((u) si(u)) ( u, Ú ). si y y y ( y ) si y y [ y si y] si y y y ( y ) si y cos y y

-TH POWER FREE ITEGERS II 7 (because lim y y si y lim y y ((si y)y), lim y y si y lim y si yy ) () si(). Substitutig this ito the above, we have {}( {}) () si() ( ) () () si(). Cosequetly, the assertio of the theorem follows at oce. REMARK. Sice, by the fuctioal equatio (s) ( s) si s () s ( s) of the Riema zeta fuctio, we see si (), ( ) ( )(si(()))() ( ) () () si(()) () () si(()) ( ). The the aearace of Theorem becomes goo as lim E [(Y () ) ] Proof of Theorem. (i) For M ( ). E [(Y () M Y () ) ] E [(Y () M ) ] E [Y () M Y () ] E [(Y () ) ]. The assertio of (i) is obvious from Claim a Theorem. (ii) By Theorem, {Y () } is L -boue, a thus for ay subsequece { i } i {i m } m : subsequece, Y ¾ L (Ç, B, ) s.t. w-limm Y () im Y.

7 T.K. DUY AD S. TAKAOBU The lim E [Y () m im Y () i ] E [Y Y () i ], ¾ Æ. But, by Claim E [Y Y () i ] ( ¾ Æ). Lettig yiels that E [Y ]. This imlies that w-lim Y (). 5. Proof of Proositio We ow tae u the roof of Proositio. Suose f ( ) satisfies the coitio (7) or (8). Fi ¾ (, ) a h ¾ C [, ] with h(). We ivie È f ()h({ }) ito three terms as (7) f ()h M( f ) h ( f () M( f ))h f ()h. To fi a limit of each term as, we reset the followig lemma: (i) Lemma. Let a b a ³ ¾ C [a, b]. Give a sequece {a }, set S(t)È t a (t ¾ Ê). The, for a y b a ³() y y S(t)³ (t) t S(y)³(y) S()³(). (ii) For a y b y ³() y ³(t) t {y} ³(y) {} y ³() {t} ³ (t) t. Proof. Let a b, ³ ¾ C [a, b] a a y b.

-TH POWER FREE ITEGERS II 73 (i) I case y, otig that a y y b, we have the left-ha sie y y y y a ³() (S() S( ))³() S()³() y y (because S(t) S(t)) y y S()³( ) S()(³() ³( )) S(y)³(y) S()³() S()³ (t) t S(y)³(y) S()³() S(t)³ (t) t S(y)³(y) S()³() S(t)³ (t) t S(y)³(y) S()³() y S(t)³ (t) t S(t)³ (t) t y S(y)³(y) S()³() y S(t)³ (t) t S(y)³(y) S()³() the right-ha sie. I case y, sice y y, the left-ha sie a ³(), y S(t)³ (t) t the right-ha sie S()(³(y) ³()) S()(³(y) ³()). Thus, we obtai the assertio of (i).

74 T.K. DUY AD S. TAKAOBU (ii) Let a ( ¾ Æ). I this case, S(t) t (t ), so by (i) y ³() y t³ (t) t y³(y) ³() y t³ (t) t y {t}³ (t) t y³(y) ³() {y}³(y){}³() [t³(t)] y y ³(t) t y {t} ³ (t) t (³(y) ³())[t³(t)]y {y}³(y){}³() ³() y y ³(t) t {y} {t} ³ (t) t. ³(y) {} REMARK 3. This ietity is calle the Euler summatio formula (cf. [3, Theorem. i Chater I]) or the Euler Maclauri summatio formula (cf. [, Lemma.]). Proof of Proositio uer the coitio (7). Æ The first term of (7). Æ - For L ¾ Æ with L, ( (L)) h L l Á l h (ote that ( (L )) L) L l l l h l (whe l, { } l. Also l l l ) L h l ( (l)) ( l) l

L l ( l) h ( (l)) ( l) -TH POWER FREE ITEGERS II 75 l t h() t l {t} h l t ( (l)) t t µ (aly Lemma (ii) for ³(t) h( t l) (( (l )) t ( l) )) L l ( l) h ( (l)) ( l) t t h() {t} h ( (l)) t t l t (ote that ( (l )) t ( l) t l) h ( (L)) t t h() L l {t} h t ( (L)) t t L h({}) L h() l µ L h ({}) µ µ l l µ (by chage of variable: t ). Æ - Let a L L() (). The L() () L(), a so L(), (L()) () (). Sice, by Æ - h ( (L())) h h ( (L())) L() L() h ({}) h({}) h() µ L() l ( (L())) h, l µ

76 T.K. DUY AD S. TAKAOBU we see L() h h ( (L())) L() µ h() l l L() h ({}) µ L() L() L() h({}) ma h() h() L() ma () h L() ma () h (ote that ma h() ma h ()) This shows that () (()) () ma () h (()) ma h () as. (8) the first term of (7) M( f ) Æ The seco term of (7). For simlicity, set a f () M( f ), S() È a. Æ - First y S(y) y y f () ym( f ) y y h({}) as. f () M( f ) {y} y M( f )

-TH POWER FREE ITEGERS II 77 y y as y. f () M( f ) M( f ) y Æ - I the same way as i Æ -, we have that for L ¾ Æ with L a h ( (L)) L a h l l ( (l)) ( l) L l ( l) ( (l)) h t l S(t) t t h()s l (aly Lemma (i) for ³(t) h( t l) (( (l )) t ( l) )) S(t) L h ( (L)) t t t h()s l l L L l h ({})S h()s l l l L h ({})S h()s. Æ -3 Fi L ¾ Æ with L. By Æ - a h ( (L)) L h ({}) S h() S ma h () L ma h () L S(( ) ) ( ) S(( ) ) ( ) S(( ) ) ( ) S(( ) ) ( ) (ote that Lµ Lµ( (L)) ( ) ( ) ) ma () h su y( (L)) S(y) y L

78 T.K. DUY AD S. TAKAOBU S(y) L y ma () h su ma h () ma h () y( (L)) su y( (L)) S(y) y su y( (L)) (L) S(y) y (L). Also ( (L)) a h ( (L)) ma h() ma h() ma h() f ()h M( f ) ( (L)) f ()M( f ) ( (L)) ( (L)) M( f ) L ma h() L M( f ) ma h() L ma h() L f () f () ( (L)) h ( (L)) f ()M( f ) ± L f () ( (L)) (L ). f () L

Combiig two estimates above, we have -TH POWER FREE ITEGERS II 79 (9) a h ma () h ma h() Ï A su y( (L)) S(y) y su y( (L)) f () S(y) (L ) y (L ) (L ) B(L ). Æ -4 Tae so that B(( )A). By Æ - y S(y) s.t. (y y ). y Let L B(( )A) ¾ Æ. For y B(( )A), B L. y y ( )A Sice ( (L )) y, su y( (L)) S(y)y. Usig this i (9), we have a h A(L ) B(L ) A A A B ( )A B ( )A B ( )A B B B B ( )A B ( )A B ( )A ( ) A B ( ) A B (( ) ( ) )A B.

7 T.K. DUY AD S. TAKAOBU Lettig yiels lim This shows that a h (( ) ( ) )A B as ². () the seco term of (7) as. 3 Æ The thir term of (7). 3 Æ - First, we chec the covergece of a series È f ()h( ). Let L, M ¾ Æ, L M. Lemma (i) for ³(t) h( t ) (L t M ) tells us that L M M L f ()h f () t M t M f () h t t h( M) M t M L L f () h( L) L L. Ê Here, otig that sice, tt, lim M M lim L L a sice h(), lim h() h (), we see the covergece of this series. 3 Æ - et, lettig L a M i the above yiels that t f ()h t f () h t t t f () h ( ) f () f () h() h() (by chage of variable: t ).

By multilyig both sies by, it turs out that -TH POWER FREE ITEGERS II 7 the thir term of (7) f () h ( ) f () h(). Thus, by the Lebesgue covergece theorem, () the thir term of (7) M( f ) M( f ) M( f ) M( f ) M( f )h ( ) M( f )h() (h ( ) ) h() ( h( )) h() [ h( )] h( ) h( (because h( ) (h( ) )( ) as ) M( f ) (by chage of variable: ) M( f ) h() 4 Æ Collectig (8), () a (), we have lim M( f ) M( f ) h({}) as. f ()h h({}) M( f ) h({}). ) h() h({}) Proof of Proositio uer the coitio (8). The argumet of Æ a 3 Æ i the revious roof is also vali i this case. Thus we have the covergeces (8) a ().

7 T.K. DUY AD S. TAKAOBU I the followig, we slightly moify the argumet of Æ i the revious roof. Let a f () M( f ), S() È a. Æ - First Clearly Æ - For L ¾ Æ with L L ( (L)) a h S(y) as y. y h ({})S Æ -3 Fi L ¾ Æ with L. By Æ - a h ( (L)) ma h () a h ( (L)) Combiig these estimates, we have () a h ma () h sua ma Ï A su y( (L)) h()s su y( (L)). S(y) (L ). y ( (L)) a h sua ma h() (L ) sua ma h() h() S(y) y su y( (L)) (L ) S(y) (L ) y (L ) C(L ). (L ).

-TH POWER FREE ITEGERS II 73 Æ -4 Tae so that C(( a choose y such that S(y)y (y y ). Let L C(( )A) ¾ Æ. For y C(( )A), su y( (L)) S(y)y. Usig this i () a h A(L ) C(L ) (( ) ( ) )A C. Lettig, a the ², we have the covergece (). Cosequetly, the assertio of Proositio uer the coitio (8) follows. Refereces [] T.K. Duy: O the istributio of -th ower free itegers, Osaa J. Math. 48 (), 7 45. [] K. Matsumoto: Riema Zeta Fuctio, Asaura, 7 (i Jaaese). [3] W. Schwarz a J. Siler: Arithmetical Fuctios, Loo Mathematical Society Lecture ote Series 84, Cambrige Uiv. Press, Cambrige, 994. [4] H. Sugita a S. Taaobu: The robability of two itegers to be co-rime, revisite o the behavior of CLT-scalig limit, Osaa J. Math. 4 (3), 945 976. Trih Khah Duy Deartmet of Mathematics Grauate School of Sciece Osaa Uiversity Osaa 56-43 Jaa e-mail: hahuy6@gmail.com Satoshi Taaobu Faculty of Mathematics a Physics Istitute of Sciece a Egieerig Kaazawa Uiversity Kaazawa 9-9 Jaa e-mail: taaob@staff.aazawa-u.ac.j