• 検索結果がありません。

Journal of Inequalities in Pure and Applied Mathematics

N/A
N/A
Protected

Academic year: 2022

シェア "Journal of Inequalities in Pure and Applied Mathematics"

Copied!
4
0
0

読み込み中.... (全文を見る)

全文

(1)

Journal of Inequalities in Pure and Applied Mathematics

http://jipam.vu.edu.au/

Volume 6, Issue 1, Article 12, 2005

HARDY-TYPE INEQUALITIES FOR HERMITE EXPANSIONS

R. BALASUBRAMANIAN AND R. RADHA THEINSTITUTE OFMATHEMATICALSCIENCES

C.I.T. CAMPUS

THARAMANI

CHENNAI- 600 113 INDIA [email protected] DEPARTMENT OFMATHEMATICS

INDIANINSTITUTE OFTECHNOLOGYMADRAS

CHENNAI- 600 036 INDIA [email protected]

Received 13 October, 2004; accepted 04 December, 2004 Communicated by H.M. Srivastava

ABSTRACT. Hardy-type inequalities are established for Hermite expansions forf Hp(R),0<

p1.

Key words and phrases: Atomic decomposition, Fourier-Hermite coefficient, Hardy spaces, Hermite functions.

2000 Mathematics Subject Classification. Primary: 42C10; Secondary: 42B30, 33C45.

1. INTRODUCTION

Hardy’s inequality for a Fourier transformF is stated as Z

R

|Ff(ξ)|p

|ξ|2−p dξ ≤CkfkpReHp 0< p≤1,

where ReHp denotes the real Hardy space consisting of the boundary values of real parts of functions in the Hardy spaceHp on the unit disc in the plane. Kanjin in [1] has proved Hardy’s inequalities for Hermite and Laguerre expansions for functions inH1. In [4] Satake has obtained Hardy’s inequalities for Laguerre expansions for Hp where 0 < p ≤ 1. In connection with regularity properties of spherical means onCn, Thangavelu [6] has proved a Hardy’s inequality for special Hermite functions. These type of inequalities for higher dimensional expansions are studied in [2], [3]. In this short note we obtain such inequalities for Hermite expansions for one dimension, namely forf ∈ Hp(R), 0< p≤ 1. In fact, it is to be noted from Theorem 2.1 that the resulting inequality for Hermite expansions(0< p≤1)is sharper than the inequalities for the classical Fourier transform as well as the Laguerre function expansion.

ISSN (electronic): 1443-5756

c 2005 Victoria University. All rights reserved.

One of the authors (R.R.) wishes to thank Prof. S. Thangavelu for initiating her into this work.

191-04

(2)

2 R. BALASUBRAMANIAN ANDR. RADHA

AnHp atom,0< p ≤1is defined to be a functionasatisfying the following conditions:

i. ais supported in an interval[b, b+h]

ii. |a(x)| ≤h−1/palmost everywhere and iii. R

Rxka(x)dx= 0for allk = 0,1,2, . . . ,h

1 p −1i

.

Making use of the atomic decomposition we define the Hardy spaceHp to be the collection of all functions f satisfying f = P

k=0λkak, whereaj is anHp - atom, λk is a sequence of complex numbers withP

k|p <∞and CkfkHp ≤X

k|p1p

≤C0kfkHp.

For various other definitions ofHp-spaces we refer to Stein [5].

2. THEMAINRESULT

LetHkdenote the Hermite polynomials Hk(x) = (−1)k dk

dxk

e−x2

ex2, k = 0,1,2, . . . .

Then the Hermite functionsh˜kare defined by

k(x) =Hk(x)e12x2, k = 0,1,2, . . . . The normalized Hermite functionshkare defined as

hk(x) = (2kk!√

π)12k(x).

These functions {hk} form an orthonormal basis for L2(R). They are eigenfunctions for the Hermite operatorH =−∆ +x2with eigenvalues2k+ 1. For more results concerning Hermite expansions, we refer to [7].

The following inequalities for Hermite functions are well known:

|hk(x)| ≤Ck121 and

d dxhk(x)

≤Ck125 .

Using these inequalities and the relation d

dxhk(x) = k

2 12

hk−1(x) +

k+ 1 2

12

hk+1(x)

we obtain the estimate

dm dxmhk(x)

≤Ck121+m2 for m= 0,1,2, . . . , which can be verified easily by applying induction onm.

Theorem 2.1. Let{hk}be the normalized Hermite functions onR. Let0 < p ≤ 1and m = h1

p

i

. Then for everyf ∈Hp(R), the Fourier - Hermite coefficient off, namely,

f(k) =ˆ Z

R

f(x)hk(x)dx, k = 0,1,2,3, . . .

satisfies the inequality

X

k=0

|fˆ(k)|p

(k+ 1)σ ≤CkfkHp, whereCis a constant andσ= 2−p12 18m+11

2m+1 = 34 +12m+61

(2−p).

J. Inequal. Pure and Appl. Math., 6(1) Art. 12, 2005 http://jipam.vu.edu.au/

(3)

HARDY-TYPEINEQUALITIESFORHERMITEEXPANSIONS 3

Proof. In order to prove the theorem, it is enough to prove that

X

k=0

|fˆ(k)|p (k+ 1)σ ≤C

for anHp-atomf. Letf be anHp atom. By considering the remainder term of the Taylor series expansion forhk(x), we write the Fourier-Hermite coefficient off as

f(k) =ˆ 1 m!

Z b+h

b

f(x)dm

dtmhk(t)(x−b)mdx,

wheret ∈[b, x]andm = h1

p

i . Then

|f(k)| ≤ˆ Chm Z b+h

b

|f(x)|

dm dtmhk(t)

dx

≤Chmk121+m2 Z b+h

b

|f(x)|dx

≤Chmk121+m2 h1p+1. Consider

X

k=0

|f(k)|ˆ p

(k+ 1)σ =X

k≤γ

|fˆ(k)|p

(k+ 1)σ +X

k>γ

|fˆ(k)|p (k+ 1)σ

=S1+S2.

We chooseγ =h−6(2m+1)6m+5 . Then

S1 ≤Chmp−1+pX

k≤γ

k

−p 12 + mp2 (k+ 1)σ.

Sinceσ = 2−p12 18m+11

2m+1 andm=h

1 p

i

, we get m

2 − 1 12

p−σ+ 1 = (6m+ 5){(m+ 1)p−1}

2m+ 1 >0.

Thus

S1 ≤Chmp−1+pγ(m2121)p−σ+1 ≤C by the choice ofγ.

On the other hand, applying Hölder’s inequality withP = 2p, we get,

S2 =X

k>γ

|f(k)|ˆ p (k+ 1)σ

≤ X

k>γ

|f(k)|ˆ 2

!p2 X

k>γ

1 (k+ 1)2−p

!2−p2

≤ kfkp2γ(2−p +1)2−p2 .

Using property (ii) of anHp-atom, we getkfkp2 ≤h−1+p2 and thus S2 ≤h−1+p2γ−σ+(2−p2 ) ≤C

J. Inequal. Pure and Appl. Math., 6(1) Art. 12, 2005 http://jipam.vu.edu.au/

(4)

4 R. BALASUBRAMANIAN ANDR. RADHA

again by the choice ofγ, thus proving our assertion.

Remark 2.2. In the case of higher dimensions, the result has been proved withσ = n4 + 12 (2−

p)(see [3]). However, here, we need an additional factor 12m+61 which approaches0asp →0.

But whenp= 1, the value ofσ= 2936, which coincides with the result of Kanjin in [1].

REFERENCES

[1] Y. KANJIN, Hardy’s inequalities for Hermite and Laguerre expansions, Bull. London Math. Soc., 29 (1997), 331–337.

[2] R. RADHA, Hardy type inequalities, Taiwanese J. Math., 4 (2000), 447–456.

[3] R. RADHA AND S. THANGAVELU, Hardy’s inequalities for Hermite and Laguerre expansions, Proc. Amer. Math. Soc., 132(12) (2004), 3525–3536.

[4] M. SATAKE, Hardy’s inequalities for Laguerre expansions, J. Math. Soc., 52(1) (2000), 17–24.

[5] E.M. STEIN, Harmonic Analysis: Real variable methods, Orthogonality and Oscillatory integrals, Princeton Univ. Press, 1993.

[6] S. THANGAVELU, On regularity of twisted spherical means and special Hermite expansion, Proc.

Ind. Acad. Sci., 103 (1993), 303–320.

[7] S. THANGAVELU, Lectures on Hermite and Laguerre expansions, Mathematical Notes, 42, Prince- ton Univ. Press, 1993.

J. Inequal. Pure and Appl. Math., 6(1) Art. 12, 2005 http://jipam.vu.edu.au/

参照

関連したドキュメント

In the paper “ Sharp inequalities for trigonometric sums in two variables,” (Illinois Journal of Mathematics, Vol.. And the results we obtained are

A NOTE ON SUMS OF POWERS WHICH HAVE A FIXED NUMBER OF PRIME FACTORS.. RAFAEL JAKIMCZUK D EPARTMENT OF

ELBERT, Integral charcterization of the principal solution of the half-linear sec- ond order differential equations, Studia Scientiarum Mathematicarum Hungarica, 36 (2000),

RUDIN, Principles of Mathematical Analysis, Third ed., Mc Graw-Hill Co., Japan, 1976. Pure

DONOGHUE, Monotone Matrix Functions and Analytic Continuation, Berlin, Heidelberg, New York, 1974.

Key words and phrases: Logarithmically convex functions, inequalities, gamma function, Riemann’s zeta function, complete elliptic integrals of the first kind.. 2000 Mathematics

By means of the convex properties of function ln Γ(x), we obtain a new proof of a generalization of a double inequality on the Euler gamma function, obtained by Jozsef Sándor..

The variant of the generalized Popovicui inequality is given in the following theorem.