• 検索結果がありません。

ACTA UNIVERSITATIS APULENSIS No 18/2009

N/A
N/A
Protected

Academic year: 2022

シェア "ACTA UNIVERSITATIS APULENSIS No 18/2009"

Copied!
10
0
0

読み込み中.... (全文を見る)

全文

(1)

SHARP FUNCTION ESTIMATE FOR MULTILINEAR COMMUTATOR OF SINGULAR INTEGRAL WITH VARIABLE

CALDER ´ON-ZYGMUND KERNEL

Zhiqiang Wang and Lanzhe Liu

Abstract: In this paper, we prove the sharp function inequality for the multilinear commutator related to the singular integral operator with variable Calder´on-Zygmund kernel. By using the sharp inequality, we obtain the Lp- norm inequality for the multilinear commutator.

2000 Mathematics Subject Classification: 42B20, 42B25.

1. Introduction

As the development of singular integral operators, their commutators have been well studied (see [1-4]). Let T be the Calder´on-Zygmund singular inte- gral operator, a classical result of Coifman, Rocherberg and Weiss (see [3]) states that commutator [b, T](f) = T(bf)−bT(f)(where b ∈ BM O(Rn)) is bounded on Lp(Rn) for 1 < p < ∞. In [6-8], the sharp estimates for some multilinear commutators of the Calder´on-Zygmund singular integral operators are obtained. The main purpose of this paper is to prove the sharp function inequality for the multilinear commutator related to the singular integral op- erator with variable Calder´on-Zygmund kernel. By using the sharp inequality, we obtain the Lp-norm inequality for the multilinear commutator.

2. Notations and Results

First let us introduce some notations (see [4][8][9]). In this paper, Q will denote a cube of Rn with sides parallel to the axes, and for a cube Q let fQ=|Q|−1RQf(x)dx and the sharp function of f is defined by

f#(x) = sup

Q3x

1

|Q|

Z

Q

|f(y)−fQ|dy.

(2)

It is well-known that (see [4]) f#(x)≈sup

Q3x c∈Cinf

1

|Q|

Z

Q

|f(y)−C|dy.

We say that b belongs to BM O(Rn) if b# belongs to L(Rn) and define

||b||BM O =||b#||L. It has been known that (see [9])

||b−b2kQ||BM O ≤Ck||b||BM O.

Let M be the Hardy-Littlewood maximal operator, that is that M(f)(x) = sup

x∈Q

|Q|−1

Z

Q

|f(y)|dy;

we write that Mp(f) = (M(|f|p))1/p for 0< p <∞.

For bj ∈BM O(Rn)(j = 1,· · ·, m), set

||~b||BM O =

m

Y

j=1

||bj||BM O.

Given some functions bj (j = 1,· · ·, m) and a positive integer m and 1 ≤j ≤ m, we denote by Cjm the family of all finite subsets σ = {σ(1),· · ·, σ(j)} of {1,· · ·, m} of j different elements. For σ ∈ Cjm, set σc ={1,· · ·, m} \σ. For

~b = (b1,· · ·, bm) and σ = {σ(1),· · ·, σ(j)} ∈ Cjm, set ~bσ = (bσ(1),· · ·, bσ(j)), bσ =bσ(1)· · ·bσ(j) and ||~bσ||BM O =||bσ(1)||BM O· · · ||bσ(j)||BM O.

In this paper, we will study some multilinear commutators as follows.

Definition 1. Let K(x) = Ω(x)/|x|n : Rn\ {0} →R. K is said to be a Calder´on-Zygmund kernel if

(a) Ω∈C(Rn\ {0});

(b) Ω is homogeneous of degree zero;

(c) RΣΩ(x)xαdσ(x) = 0 for all multi-indices α ∈ (N ∪ {0})n with |α| = N, where Σ ={x∈Rn : |x|= 1} is the unit sphere of Rn.

Definition 2. Let K(x, y) = Ω(x, y)/|y|n : Rn×(Rn\ {0}) → R. K is said to be a variable Calder´on-Zygmund kernel if

(d) K(x,·) is a Calder´on-Zygmund kernel for a.e. x∈Rn; (e) max|γ|≤2n

|γ|

γyΩ(x, y)

L(Rn×Σ) =M < ∞.

(3)

Suppose bj (j = 1,· · ·, m) are the fixed locally integrable functions on Rn. LetT be the singular integral operator with variable Calder´on-Zygmund kernel as

T(f)(x) =

Z

Rn

K(x, x−y)f(y)d(y),

whereK(x, x−y) = Ω(x, x−y)

|x−y|n and that Ω(x, y)/|y|n is a variable Calder´on- Zygmund kernel. The multilinear commutator of singular integral with variable Calder´on-Zygmund kernel is defined by

T~b(f)(x) =

Z

Rn m

Y

j=1

(bj(x)−bj(y))K(x, x−y)f(y)dy.

Note that whenb1 =···=bm,T~bis just themorder commutator (see [1][5]).

It is well known that commutators are of great interest in harmonic analysis and have been widely studied by many authors (see [1-3][5-8]). Our main purpose is to establish the sharp inequality for the multilinear commutator.

Now we state our theorems as following.

Theorem 1. Let bj ∈ BM O(Rn) for j = 1,· · ·, m. Then for any 1 <

r < ∞, there exists a constant C >0 such that for any f ∈C0(Rn) and any x∈Rn,

(T~b(f))#(x)≤C||~b||BM O

Mr(f)(x) +

m

X

j=1

X

σ∈Cjm

Mr(T~b

σc(f))(x)

.

Theorem 2. Let bj ∈BM O(Rn) for j = 1,· · ·, m. Then T~b is bounded on Lp(Rn) for 1< p <∞.

3. Proof of Theorem To prove the theorems, we need the following lemmas.

Lemma 1. (see[10])Let1< p < ∞andT be the singular integral operator with variable Calder´on-Zygmund kernel. Then T is bounded on Lp(Rn).

Lemma 2. Let 1< r <∞, bj ∈BM O(Rn) for j = 1,· · ·, k. Then 1

|Q|

Z

Q k

Y

j=1

|bj(y)−(bj)Q|dy≤C

k

Y

j=1

||bj||BM O

(4)

and

1

|Q|

Z

Q k

Y

j=1

|bj(y)−(bj)Q|rdy

1/r

≤C

k

Y

j=1

||bj||BM O.

Proof. Choose 1< pj <∞j = 1,· · ·, ksuch that 1/p1+· · ·+ 1/pk = 1, we obtain, by H¨older’s inequality,

1

|Q|

Z

Q k

Y

j=1

|bj(y)−(bj)Q|dy≤

k

Y

j=1

1

|Q|

Z

Q

|bj(y)−(bj)Q|pjdy

!1/pj

≤C

k

Y

j=1

||bj||BM O

and

1

|Q|

Z

Q k

Y

j=1

|bj(y)−(bj)Q|rdy

1/r

k

Y

j=1

1

|Q|

Z

Q

|bj(y)−(bj)Q|pjrdy

!1/pjr

C

k

Y

j=1

||bj||BM O.

The lemma follows.

Proof of Theorem 1. It suffices to prove forf ∈C0(Rn) and some constant C0, the following inequality holds:

1

|Q|

Z

Q

|T~b(f)(x)−C0|dx≤C||~b||BM O

Mr(f)(˜x) +

m

X

j=1

X

σ∈Cjm

Mr(T~b

σc(f)(˜x))

.

Fix a cube Q=Q(x0, r) and ˜x∈Q.

We first consider the Case m=1 . Write, forf1 =f χ2Q and f2 =f χ(2Q)c, Tb1(f)(x) = (b1(x)−(b1)2Q)T(f)(x)−T((b1−(b1)2Q)f1)(x)−T((b1−(b1)2Q)f2)(x).

LetC0 =T(((b1)2Q−b1)f2)(x0), then

|Tb1(f)(x)−C0|

≤ |(b1(x)−(b1)2Q)T(f)(x) +T((b1)2Q−(b1)f1)(x) +T(((b1)2Q−b1)f2)(x)−T(((b1)2Q−b1)f2)(x0)|

≤ |(b1(x)−(b1)2Q)T(f)(x)|+|T((b1)2Q−(b1)f1)(x)|

+|T(((b1)2Q−b1)f2)(x)−T(((b1)2Q−b1)f2)(x0)|

= A(x) +B(x) +C(x).

(5)

ForA(x), by H¨older’s inequality with exponent 1/r+ 1/r0 = 1, we get 1

|Q|

Z

Q

A(x)dx

= 1

|Q|

Z

Q

|b1(x)−(b1)2Q||T(f)(x)|dx

≤ C 1

|2Q|

Z

2Q

|b1(x)−(b1)2Q|r0dx

!1/r0

1

|Q|

Z

Q

|T(f)(x)|rdx

!1/r

≤ C||b1||BM OMr(T(f))(˜x).

For B(x), choose p such that 1 < p < r, and 1 < q < ∞, pq = r, by the boundedness of T onLp(Rn) and the H¨older’s inequality, we obtain

1

|Q|

Z

Q

B(x)dx

= 1

|Q|

Z

Q

T((b1−(b1)2Q)f1)(x)dx

≤ 1

|Q|

Z

Rn

[T((b1−(b1)2Q)f χ2Q)(x)]pdx

!1/p

≤ C 1

|Q|

Z

Rn

(|b1(x)−(b1)2Q||f(x)χ2Q(x)|)pdx

!1/p

≤ C 1

|Q|1/p

Z

2Q

|b1(x)−(b1)2Q|1/pq0dx

1/pq0Z

2Q

|f(x)|pqdx

1/pq

≤ C 1

|2Q|

Z

2Q

|f(x)|rdx

!1/r

1

|2Q|

Z

2Q

|b1(x)−(b1)2Q|pq0dx

!1/pq0

≤ C||b1||BM OMr(f)(˜x).

ForC(x), by [11], we know that T~b(f)(x) =

X

k=1 gk

X

h=1

ahk(x)

Z

Rn

Yhk(x−y)

|x−y|n+m

m

Y

j=1

(bj(x)−bj(y))f(y)dy

where gk ≤Ckn−2, ||ahk||L ≤Ck−2n, |Yhk(x−y)| ≤Ckn/2−1 and

Yhk(x−y)

|x−y|n −Yhk(x0−y)

|x0 −y|n

≤Ckn/2|x−x0|/|x0−y|n+1

(6)

for |x−y|>2|x0−x|>0. So we get, by Minkowski’s inequality and H¨older’s inequality,

C(x) =

Z

Rn

(K(x, x−y)−K(x0, x0−y))((b1)2Q−b1(y))f2(y)dy

≤C

Z

(2Q)c

Ω(x, x−y)

|x−y|n − Ω(x0, x0−y)

|x0−y|n

|(b1)2Q−b1(y)||f(y)|dy

≤C

X

l=1

Z

2l+1Q\2lQ

X

k=1 gk

X

h=1

|ahk(x)|

Z

Rn

Yhk(x−y)

|x−y|n − Yhk(x0−y)

|x0−y|n

|(b1)2Q−b1(y)|·

·|f(y)|dy≤C

X

k=1

k−2n·kn/2

X

l=1

Z

2l+1Q\2lQ

|x−x0|

|x0−y|n+1|b1(y)−(b1)2Q||f(y)|dy

≤C

X

k=1

k−3n/2

X

l=1

r (2lr)n+1

Z

2l+1Q

|b1(y)−(b1)2Q||f(y)|dy

≤C

X

l=1

2−l 1

|2l+1Q|

Z

2l+1Q

|b1(y)−(b1)2Q|r0dy

!1/r0

1

|2l+1Q|

Z

2l+1Q

|f(y)|rdy

!1/r

≤C

X

l=1

l2−l||b1||BM OMr(f)(˜x)≤C||b1||BM OMr(f)(˜x),

thus 1

|Q|

Z

Q

C(x)dx ≤C||b1||BM OMr(f)(˜x).

Now, we consider theCase m ≥2, we have known that, forb= (b1,· · ·, bm),

T~b(f)(x) =

Z

Rn

m

Y

j=1

(bj(x)−bj(y))

K(x, x−y)f(y)dy

=

Z

Rn m

Y

j=1

[(bj(x)−(bj)2Q)−(bj(y)−(bj)2Q)]K(x, x−y)f(y)dy

=

m

X

j=0

X

σ∈Cmj

(−1)m−j(b(x)−(b)2Q)σ

Z

Rn

(b(y)−(b)2Q)σK(x, x−y)f(y)dy

= (b1(x)−(b1)2Q)· · ·(bm(x)−(bm)2Q)T(f)(x) +(−1)mT((b1−(b1)2Q)· · ·(bm−(bm)2Q)f)(x)

(7)

+

m−1

X

j=1

X

σ∈Cmj

(−1)m−j(b(x)−(b)2Q)σ

Z

Rn

(b(y)−b(x))σcK(x, x−y)f(y)dy

= (b1(x)−(b1)2Q)· · ·(bm(x)−(bm)2Q)T(f)(x) +(−1)mT((b1−(b1)2Q)· · ·(bm−(bm)2Q)f)(x) +

m−1

X

j=1

X

σ∈Cmj

(−1)m−j(b(x)−(b)2Q)σT~b

σc(f)(x), thus,

|T~b(f)(x)−T((b1−(b1)2B)· · ·(bm−(bm)2B)f2)(x0)|

≤ |(b1(x)−(b1)2Q)· · ·(bm(x)−(bm)2Q)T(f)(x)|

+

m−1

X

j=1

X

σ∈Cjm

|(b(x)−(b)2Q)σT~b

σc(f)(x)|

+|T((b1−(b1)2Q)· · ·(bm−(bm)2Q)f1)(x)|+

|T((b1−(b1)2Q)· · ·(bm−(bm)2Q)f2)(x)−T((b1−(b1)2Q)· · ·(bm−(bm)2Q)f2)(x0)|

=I1(x) +I2(x) +I3(x) +I4(x).

ForI1(x), by H¨older’s inequality with exponent 1/p1+· · ·+1/pm+1/r= 1, where 1< pj <∞, j = 1,· · ·, m, we get

1

|Q|

Z

Q

I1(x)dx

≤ 1

|Q|

Z

Q

|b1(x)−(b1)2Q| · · · |bm(x)−(bm)2Q||T(f)(x)|dx

m

Y

j=1

1

|Q|

Z

Q

|bj(x)−(bj)2Q|pjdx

!1/pj

1

|Q|

Z

Q

|T(f)(x)|rdx

!1/r

≤ C||~b||BM OMr(T(f))(˜x).

ForI2(x), by the Minkowski’s and H¨older’s inequality, we get 1

|Q|

Z

Q

I2(x)dx

m−1

X

j=1

X

σ∈Cmj

1

|Q|

Z

Q

|(b(x)−(b)2Q)σ||T~b

σc(f)(x)|dx

(8)

≤C

m−1

X

j=1

X

σ∈Cjm

1

|2Q|

Z

2Q

|(b(x)−(b)2Q)σ|r0dµ(x)

!1/r0

1

|Q|

Z

Q

|T~b

σc(f)(x)|rdx

!1/r

≤C

m−1

X

j=1

X

σ∈Cjm

||~bσ||BM OMr(T~b

σc(f))(˜x).

For I3(x), choose 1 < p < r, 1 < qj < ∞, j = 1,· · ·, m such that 1/q1+· · ·+ 1/qm+p/r= 1, by the boundedness ofT onLp(Rn) and H¨older’s inequality, we get

1

|Q|

Z

Q

I3(x)dx

≤ 1

|Q|

Z

Rn

|T((b1−(b1)2Q)· · ·(bm−(bm)2Q)f χ2Q)(x)|pdx

!1/p

≤ C 1

|Q|

Z

Rn

|b1(x)−(b1)2Q|p· · · |bm(x)−(bm)2Q|p|f(x)χ2Q(x)|pdx

!1/p

≤ C 1

|2Q|

Z

2Q

|f(x)|rdx

!1/r m

Y

j=1

1

|2Q|

Z

2Q

|bj(x)−(bj)2B|pqjdx

!1/pqj

≤ C||~b||BM OMr(f)(˜x).

ForI4(x), choose 1< pj <∞j = 1,· · ·, msuch that 1/p1+· · ·+1/pm+1/r= 1, by Minkowski’s inequality and H¨older’s inequality, we obtain

I4(x)≤C

Z

(2Q)c

Ω(x, x−y)

|x−y|n −Ω(x0, x0−y)

|x0−y|n

|

m

Y

j=1

(bj(y)−(bj)2Q)||f(y)|dy

≤C

X

l=1

Z

2l+1Q\2lQ

" X

k=1 gk

X

h=1

|ahk(x)|

Z

Rn

Yhk(x−y)

|x−y|n −Yhk(x0−y)

|x0 −y|n

#

×|

m

Y

j=1

(bj(y)−(bj)2Q)||f(y)|dy

≤C

X

k=1

k−2n·kn/2

X

l=1

Z

2l+1Q\2lQ

|x−x0|

|x0−y|n+1|

m

Y

j=1

(bj(y)−(bj)2Q)||f(y)|dy

≤C

X

k=1

k−3n/2

X

l=1

r (2lr)n+1

Z

2l+1Q

|

m

Y

j=1

(bj(y)−(bj)2Q)||f(y)|dy

(9)

≤C

X

l=1

2−l 1

|2l+1Q|

Z

2l+1Q

|f(y)|rdy

!1/r

·

·

m

Y

j=1

1

|2l+1Q|

Z

2l+1Q

|bj(y)−(bj)2Q|pjdy

!1/pj

≤C

X

l=1

lm2−l

m

Y

j=1

||bj||BM OMr(f)(˜x)≤C||~b||BM OMr(f)(˜x),

thus 1

|Q|

Z

Q

I4(x)dx≤C||~b||BM OMr(f)(˜x).

This completes the proof of the theorem.

Proof of Theorem 2. Choose 1< r < p in Theorem 1. We first consider the case m=1, we have

||Tb1(f)||Lp ≤ ||M(Tb1(f))||Lp ≤Ck(Tb1(f))#kLp

≤ CkMr(T(f))kLp +CkMr(f)kLp

≤ CkT(f)kLp+CkMr(f)kLp

≤ CkfkLp+CkfkLp

≤ CkfkLp.

When m ≥ 2, we may get the conclusion of Theorem 2 by induction. This finishes the proof.

References

[1] A. P. Calder´on and A. Zygmund, On singular integrals with variable kernels, Appl. Anal., 7, (1978), 221-238.

[2] F. Chiarenza, M. Frasca and P. Longo,InteriorW2,p-estimates for non- divergence elliptic equations with discontinuous coefficients, Ricerche Mat., 40, (1991), 149-168.

[3] R. Coifman, R. Rochberg and G. Weiss, Factorization theorems for Hardy spaces in several variables, Ann. of Math., 103, (1976), 611-635.

[4] G. Di Fazio and M. A. Ragusa, Interior estimates in Morrey spaces for strong solutions to nondivergence form equations with discontinuous coef- ficients, J. Func. Anal., 112, (1993), 241-256.

(10)

[5] J. Garcia-Cuerva and J.L.Rubio de Francia, Weighted norm inequalities and related topics, North-Holland Math., 16, Amsterdam, 1985.

[6] L. Z. Liu,The continuity for multilinear singular integral operators with variable Calder´on-Zygmund kernel on Hardy and Herz spaces, Siberia Elec- tronic Math. Reports, 2, (2005), 156-166.

[7] L. Z. Liu, Good λ estimate for multilinear singular integral operators with variable Calder´on-Zygmund kernel, Kragujevac J. of Math., 27, (2005), 19-30.

[8] L. Z. Liu, Weighted estimates of multilinear singular integral operators with variable Calder´on-Zygmund kernel for the extreme cases, Vietnam J. of Math., 34, (2006), 51-61.

[9] S. Z. Lu, D. C. Yang and Z. S. Zhou, Oscillatory singular integral op- erators with Calder´on-Zygmund kernels, Southeast Asian Bull. of Math., 23, (1999), 457-470.

[10] C. P´erez, Endpoint estimate for commutators of singular integral op- erators, J. Func. Anal., 128, (1995), 163-185.

[11] C. P´erez and G. Pradolini, Sharp weighted endpoint estimates for com- mutators of singular integral operators, Michigan Math. J., 49, (2001), 23-37.

[12] C. P´erez and R.Trujillo-Gonzalez, Sharp Weighted estimates for mul- tilinear commutators, J. London Math. Soc, 65, (2002), 672-692.

[13] E. M. Stein, Harmonic Analysis: real variable methods, orthogonality and oscillatory integrals, Princeton Univ. Press, Princeton NJ, 1993.

[14] A. Torchinsky, Real variable methods in harmonic analysis, Pure and Applied Math., 123, Academic Press, New York, 1986.

[15] H. Xu and L. Z. Liu., Weighted boundedness for multilinear singular integral operator with variable Calder´on-Zygmund kernel, African Diaspora J.

of Math., 6, (2008), 1-12.

Authors:

Zhiqiang Wang and Lanzhe Liu College of Mathematics

Changsha University of Science and Technology Changsha, 410077, P.R.of China

E-mail: [email protected]

参照

関連したドキュメント

The steps of specification of the econometric model by using formalization and choosing the form of functions which describe the model, the selection and measurement of variables in

The purpose of this contribution is to give some common fixed point theorems for occasionally weakly compatible mappings satisfying an altering distance in a metric space..

The degree of sensitivity can range from no change of the optimal solution, with the only change in the optimal value of the objective function, up to a considerable change of

In many cases is very useful to know how probable is that the aggregate risk belongs to a certain neighbourhood of its expected value (as in the previous example).. Another field

Y. Chen defined and studied slant immersions in complex manifolds. Fernandez [2] studied and characterized slant submanifolds in the case of K-contact and Sasaki manifolds. The

Making use of certain linear operator, we define a new subclass of meromorphically uniformly convex functions with positive coefficients and obtain coefficient estimates, growth

We construct new Yang-Baxter systems from algebra and bialgebra

Abstract: We prove weighted inequalities for commutators of one-sided singular integrals (given by a Calder´on-Zygmund kernel with support in (−∞, 0)) with BMO functions.. We give