Skip to main content

Weighted endpoint estimates for multilinear commutator of singular integral operators with non-smooth kernels

Abstract

In this paper, we prove the weighted endpoint estimates for multilinear commutator of singular integral operators with non-smooth kernels.

1 Introduction

Let bBMO( R n ) and T be the Calderón-Zygmund operator, the commutator [b,T] generated by b and T is defined by [b,T](f)(x)=b(x)T(f)(x)T(bf)(x). A classical result of Coifman et al. (see [1]) proved that the commutator [b,T] is bounded on L p ( R n ) (1<p<). In [2, 3], the boundedness properties of the commutators for the extreme values of p are obtained. In this paper, we will introduce the multilinear commutator of singular integral operators with non-smooth kernels and prove the weighted boundedness properties of the operator for the extreme cases.

First let us introduce some notations (see [312]). In this paper, Q will denote a cube of R n with sides parallel to the axes. For a cube Q and a function b, let b Q = | Q | 1 Q b(x)dx and b(Q)= Q b(x)dx, the sharp function of b is defined by

b # (x)= sup Q x 1 | Q | Q |b(y) b Q |dy.

It is well known that (see [6])

b # (x)= sup Q x inf c C 1 | Q | Q |b(y)c|dy.

Moreover, for a weight function ω (that is, a non-negative locally integrable function), b is said to belong to BMO(ω) if b # L (ω) and define b BMO ( ω ) = b # L ( ω ) , if ω=1, we denote BMO(ω)=BMO( R n ). It is well known that (see [11])

b b 2 k Q BMO Ck b BMO .

The A p weight is defined by (see [6])

A p = { 0 < ω L loc 1 ( R n ) : sup Q ( 1 | Q | Q ω ( x ) d x ) ( 1 | Q | Q ω ( x ) 1 / ( p 1 ) d x ) p 1 < } , 1 < p <

and

A 1 = { 0 < ω L loc 1 ( R n ) : sup Q x 1 | Q | Q ω ( y ) d y c ω ( x ) , a.e. } .

Definition 1 A family of operators D t , t>0, is said to be an ‘approximation to the identity’ if, for every t>0, D t can be represented by the kernel a t (x,y) in the following sense:

D t (f)(x)= R n a t (x,y)f(y)dy

for every f L p ( R n ) with p1, and a t (x,y) satisfies

| a t (x,y)| h t (x,y)=C t n / 2 s ( | x y | 2 / t ) ,

where s is a positive, bounded, and decreasing function satisfying

lim r r n + ϵ s ( r 2 ) =0

for some ϵ>0.

Definition 2 A linear operator T is called the singular integral operators with non-smooth kernels if T is bounded on L 2 ( R n ) and associated with a kernel K(x,y) such that

T(f)(x)= R n K(x,y)f(y)dy

for every continuous function f with compact support, and for almost all x not in the support of f.

  1. (1)

    There exists an ‘approximation to the identity’ { B t ,t>0} such that T B t has associated kernel k t (x,y) and there exist c 1 , c 2 >0 so that

    | x y | > c 1 t 1 / 2 |K(x,y) k t (x,y)|dx c 2 for all y R n .
  2. (2)

    There exists an ‘approximation to the identity’ { A t ,t>0} such that A t T has associated kernel K t (x,y) which satisfies

    | K t (x,y)| c 4 t n / 2 if |xy| c 3 t 1 / 2

    and

    |K(x,y) K t (x,y)| c 4 t δ / 2 | x y | n δ if |xy| c 3 t 1 / 2

    for some c 3 , c 4 >0, δ>0.

Given some locally integrable functions b j (j=1,,m). The multilinear operator associated to T is defined by

T b (f)(x)= R n [ j = 1 m ( b j ( x ) b j ( y ) ) ] K(x,y)f(y)dy.

Definition 3 Given the ‘approximations to the identity’ { A t ,t>0} and a weight function ω.

  1. (1)

    The weighted BMO space associated with { A t ,t>0} is defined by

    BMO A (ω)= { f L loc 1 ( R n ) : f BMO A ( ω ) < } ,

    where

    f BMO A ( ω ) = sup Q 1 ω ( Q ) Q |f(x) A t Q (f)(x)|ω(x)dx,

    t Q =l ( Q ) 2 and l(Q) denotes the side length of Q.

  2. (2)

    The weighted central BMO space associated with { A t ,t>0} is defined by

    CMO A (ω)= { f L loc 1 ( R n ) : f CMO A ( ω ) < } ,

    where

    f CMO ( ω ) = sup r > 1 1 ω ( Q ( 0 , r ) ) Q |f(x) A t Q f(x)|ω(x)dx,

    and t Q = r 2 .

Definition 4 Let 1<p< and ω be a weighted function on R n . We shall call B p (ω) the space of those functions f on R n , such that

f B p ( ω ) = sup r > 1 [ ω ( Q ( 0 , r ) ) ] 1 / p f χ Q ( 0 , r ) L p ( ω ) <.

For b j BMO( R n ) (j=1,,m), set b BMO = j = 1 m b j BMO . Given a positive integer m and 1jm, we denote by C j m the family of all finite subsets σ={σ(1),,σ(j)} of {1,,m} of j different elements. For σ C j m , set σ c ={1,,m}σ. For b =( b 1 ,, b m ) and σ={σ(1),,σ(j)} C j m , set b σ =( b σ ( 1 ) ,, b σ ( j ) ), b σ = b σ ( 1 ) b σ ( j ) and b σ BMO = b σ ( 1 ) BMO b σ ( j ) BMO .

2 Theorems and proofs

We begin with some preliminaries lemmas.

Lemma 1 ([5, 7])

Let ω A 1 , 1<p, and T be the singular integral operators with non-smooth kernels. Then T is boundedness on L p (w).

Lemma 2 Let ω A 1 , { A t ,t>0} be anapproximation to the identityand bBMO( R n ). Then

  1. (a)

    for every f L ( R n ), 1p<, and any cube Q,

    ( 1 | Q | Q | A t Q ( ( b b Q ) f ) ( y ) | p d y ) 1 / p C b BMO f L ;
  2. (b)

    for every f B p (ω), 1r<p<, and any cube Q,

    ( 1 ω ( Q ) Q | A t Q ( ( b b Q ) f ) ( y ) | r ω ( y ) d y ) 1 / r C b BMO f B p ( ω ) ,

where t Q =l ( Q ) 2 and l(Q) denotes the side length of Q.

Proof (a) Write

( 1 | Q | Q | A t Q ( ( b b Q ) f ) ( y ) | p d y ) 1 / p ( 1 | Q | Q R n h t Q ( x , y ) p | ( b ( y ) b Q ) f ( y ) | p d y d x ) 1 / p ( 1 | Q | Q 2 Q h t Q ( x , y ) p | ( b ( y ) b Q ) f ( y ) | p d y d x ) 1 / p + ( k = 1 1 | Q | Q 2 k + 1 Q 2 k Q h t Q ( x , y ) p | ( b ( y ) b Q ) f ( y ) | p d y d x ) 1 / p = I 1 + I 2 .

We have, by Hölder’s inequality,

I 1 ( C | Q | | 2 Q | Q 2 Q | ( b ( y ) b Q ) f ( y ) | p d y d x ) 1 / p C f L ( 1 | 2 Q | 2 Q | b ( y ) b Q | p d y ) 1 / p C b BMO f L .

For I 2 , for xQ and y 2 k + 1 Q 2 k Q, we have |xy| 2 k 1 t Q and h t Q (x,y)C s ( 2 2 ( k 1 ) ) | Q | . Thus

I 2 C k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( 1 | 2 k + 1 Q | 2 k + 1 Q | ( b ( y ) b Q ) f ( y ) | p d y ) 1 / p C f L k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( 1 | 2 k + 1 Q | 2 k + 1 Q | b ( y ) b 2 k + 1 Q | p d y ) 1 / p + C f L k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) | b Q b 2 k + 1 Q | C f L k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( k + 1 ) b BMO C b BMO f L ,

where the last inequality follows from

k = 2 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) (k+1)C k = 2 k 2 ( k 1 ) ϵ <

for some ϵ>0.

  1. (b)

    Write

    ( 1 ω ( Q ) Q | A t Q ( ( b b Q ) f ) ( y ) | r ω ( y ) d y ) 1 / r ( 1 ω ( Q ) Q R n h t Q ( x , y ) r | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y d x ) 1 / r ( 1 ω ( Q ) Q 2 Q h t Q ( x , y ) p | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y d x ) 1 / r + ( k = 1 1 ω ( Q ) Q 2 k + 1 Q 2 k Q h t Q ( x , y ) r | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y d x ) 1 / r = I + II .

For I, since ω A 1 , ω satisfies the reverse of Hölder’s inequality

( 1 | Q | Q ω ( x ) q d x ) 1 / q C | Q | Q ω(x)dx

for some 1<q<, and ω ( Q 2 ) | Q 2 | | Q 1 | ω ( Q 1 ) C for all cubes Q 1 , Q 2 with Q 1 Q 2 , ω A p / r u for 1<u,v< with u v=q and p>ru (see [6]). We have, by Hölder’s inequality,

I ( C ω ( Q ) | Q | Q 2 Q | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y d x ) 1 / r I C ( 1 ω ( Q ) 2 Q | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y ) 1 / r I C [ | 2 Q | ω ( Q ) ( 1 | 2 Q | 2 Q | b ( y ) b Q | r u ω ( y ) u d y ) 1 / u ( 1 | 2 Q | 2 Q | f ( y ) | r u d y ) 1 / u ] 1 / r I C ( | 2 Q | ω ( Q ) ) 1 / r ( 1 | 2 Q | 2 Q | b ( y ) b Q | r u v d y ) 1 / r u v ( 1 | 2 Q | 2 Q ω ( y ) u v d y ) 1 / r u v I × ( 1 | 2 Q | 2 Q | f ( y ) | r u d y ) 1 / r u I C b BMO ( | 2 Q | ω ( 2 Q ) ) 1 / r ( ω ( 2 Q ) | 2 Q | ) 1 / r ( 1 | 2 Q | 2 Q | f ( y ) | r u ω ( y ) r u p ω ( y ) r u p d y ) 1 / r u I C b BMO ( 1 | 2 Q | 2 Q ( | f ( y ) | r u ω ( y ) r u p ) p r u d y ) 1 / p ( 1 | 2 Q | 2 Q ω ( y ) r u p p p r u d y ) ( p r u ) / p r u I C b BMO ( 1 | 2 Q | ) 1 / p f χ 2 Q L p ( ω ) ( 1 | 2 Q | 2 Q ω ( y ) d y ) 1 / p I × [ ( 1 | 2 Q | 2 Q ω ( y ) d y ) ( 1 | 2 Q | 2 Q ω ( y ) 1 p r u 1 d y ) p r u 1 ] 1 / p I C b BMO ω ( 2 Q ) 1 / p f χ 2 Q L p ( ω ) I C b BMO f B p ( ω ) ; II C ( | Q | ω ( Q ) ) 1 / r k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( 1 | 2 k + 1 Q | 2 k + 1 Q | ( b ( y ) b Q ) f ( y ) | r ω ( y ) d y ) 1 / r II C ( | Q | ω ( Q ) ) 1 / r k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( 1 | 2 k + 1 Q | 2 k + 1 Q | b ( y ) b 2 k + 1 Q | r u ω ( y ) u d y ) 1 / r u II × ( 1 | 2 k + 1 Q | 2 k + 1 Q | f ( y ) | r u d y ) 1 / r u II C ( | Q | ω ( Q ) ) 1 / r k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( 1 | 2 k + 1 Q | 2 k + 1 Q | b ( y ) b 2 k + 1 Q | r u v d y ) 1 / r u v II × ( 1 | 2 k + 1 Q | 2 k + 1 Q ω ( y ) u v d y ) 1 / r u v ( 1 | 2 k + 1 Q | 2 k + 1 Q f ( y ) r u d y ) 1 / r u II C b BMO k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( k + 1 ) ( | Q | ω ( Q ) ω ( 2 k + 1 Q ) | 2 k + 1 Q | ) 1 / r II × ( 1 | 2 k + 1 Q | 2 k + 1 Q f ( y ) r u d y ) 1 / r u II C k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) ( k + 1 ) b BMO ω ( 2 k + 1 Q ) 1 / p f χ 2 k + 1 Q L p ( ω ) II C b BMO f B p ( w ) .

This completes the proof. □

Theorem 1 Let T be the singular integral operators with non-smooth kernels, ω A 1 and b =( b 1 ,, b m ) with b j BMO( R n ) for 1jm. Then T b is bounded from L (ω) to BMO A (ω).

Proof It suffices to prove, for f C 0 ( R n ), the following inequality holds:

1 ω ( Q ) Q | T b (f)(x) A t Q T b (f)(x)|ω(x)dxC f L ( ω ) .

We fix a cube Q=Q( x 0 ,d). We decompose f into f= f 1 + f 2 with f 1 =f χ Q , f 2 =f χ ( R n Q ) .

When m=1, set ( b 1 ) Q = | Q | 1 Q b 1 (y)dy, we have

T b 1 ( f ) ( x ) = R n [ ( b 1 ( x ) ( b 1 ) Q ) ( b 1 ( y ) ( b 1 ) Q ) ] K ( x , y ) f ( y ) d y = ( b 1 ( x ) ( b 1 ) Q ) R n K ( x , y ) f ( y ) d y R n ( b 1 ( y ) ( b 1 ) Q ) K ( x , y ) f ( y ) d y

and

A t Q T b 1 (f)(x)= ( b 1 ( x ) ( b 1 ) Q ) R n K t (x,y)f(y)dy R n ( b 1 ( y ) ( b 1 ) Q ) K t (x,y)f(y)dy.

Then

| T b 1 ( f ) ( x ) A t Q T b 1 ( f ) ( x ) | | ( b 1 ( x ) ( b 1 ) Q ) R n K ( x , y ) f ( y ) d y | + | R n ( b 1 ( y ) ( b 1 ) Q ) K ( x , y ) f 1 ( y ) d y | + | ( b 1 ( x ) ( b 1 ) Q ) R n K t ( x , y ) f ( y ) d y | + | R n ( b 1 ( y ) ( b 1 ) Q ) K t ( x , y ) f 1 ( y ) d y | + | R n ( b 1 ( y ) ( b 1 ) Q ) ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | = I 1 ( x ) + I 2 ( x ) + I 3 ( x ) + I 4 ( x ) + I 5 ( x ) .

For I 1 (x), let 1/p+1/ p =1, 1/q+1/ q =1, by the reverse of Hölder’s inequality with 1<q<, Lemma 1, and Hölder’s inequality, we have

1 ω ( Q ) Q | I 1 ( x ) | ω ( x ) d x C ω ( Q ) ( Q | b 1 ( x ) ( b 1 ) Q | p ω ( x ) d x ) 1 / p ( R n | T ( f ) ( x ) | p ω ( x ) χ Q ( x ) d x ) 1 / p C ω ( Q ) ( Q | b 1 ( x ) ( b 1 ) Q | p ω ( x ) d x ) 1 / p ( R n | f ( x ) | p ω ( x ) χ Q ( x ) d x ) 1 / p C ω ( Q ) ( Q | b 1 ( x ) ( b 1 ) Q | p ω ( x ) d x ) 1 / p f L ( ω ) ( Q ω ( x ) d x ) 1 / p C ω ( Q ) [ ( Q | b 1 ( x ) ( b 1 ) Q | p q d x ) 1 / q ( Q ω ( x ) q d x ) 1 / q ] 1 / p f L ( ω ) ω ( Q ) 1 / p C ω ( Q ) 1 / p 1 | Q | 1 / p b 1 BMO ( 1 | Q | Q ω ( x ) q d x ) 1 / p q f L ( ω ) C b 1 BMO f L ( ω ) .

For I 2 (x), taking p>1, by Hölder’s inequality, we have

1 ω ( Q ) Q | I 2 ( x ) | ω ( x ) d x ( 1 ω ( Q ) R n | T ( ( b 1 ( b 1 ) Q ) f 1 ) ( x ) | p ω ( x ) χ Q ( x ) d x ) 1 / p C ω ( Q ) 1 / p ( R n | ( b 1 ( x ) ( b 1 ) Q ) f 1 ( x ) | p ω ( x ) χ Q ( x ) d x ) 1 / p C ω ( Q ) 1 / p [ ( Q | b 1 ( x ) ( b 1 ) Q | p q d x ) 1 / q ( Q | f ( x ) | p q ω ( x ) q d x ) 1 / q ] 1 / p C ω ( Q ) 1 / p ( Q | b 1 ( x ) ( b 1 ) Q | p q d x ) 1 / p q ( Q | f ( x ) | p q ω ( x ) q d x ) 1 / p q C ω ( Q ) 1 / p ( Q | b 1 ( x ) ( b 1 ) Q | p q d x ) 1 / p q ( Q ω ( x ) q d x ) 1 / p q f L ( ω ) C ω ( Q ) 1 / p | Q | 1 / p q b 1 BMO | Q | 1 / p q ( 1 | Q | Q ω ( x ) q d x ) 1 / p q f L ( ω ) C b 1 BMO ( | Q | ω ( Q ) ) 1 / p ( 1 | Q | Q ω ( x ) d x ) 1 / p f L ( ω ) C b 1 BMO f L ( ω ) .

For I 3 (x) and I 4 (x), we get, for 1< p 1 , p 2 < with 1/ p 1 +1/ p 2 +1/q=1,

1 ω ( Q ) Q | I 3 ( x ) | ω ( x ) d x C ω ( Q ) Q | b 1 ( x ) ( b 1 ) Q | | A t Q ( f ) ( x ) | ω ( x ) d x C | Q | ω ( Q ) ( 1 | Q | Q | b 1 ( x ) ( b 1 ) Q | p 1 d x ) 1 / p 1 × ( 1 | Q | Q | A t Q ( f ) ( x ) | p 2 d x ) 1 / p 2 ( 1 | Q | Q ω ( x ) q d x ) 1 / q C | Q | ω ( Q ) b 1 BMO f L ( ω ) ω ( Q ) | Q | C b 1 BMO f L ( ω ) , 1 ω ( Q ) Q | I 4 ( x ) | ω ( x ) d x 1 ω ( Q ) R n | A t Q ( ( b 1 ( b 1 ) Q ) f 1 ) ( x ) | ω ( x ) d x C | Q | ω ( Q ) ( 1 | Q | Q | A t Q ( ( b 1 ( b 1 ) Q ) f 1 ) ( x ) | q d x ) 1 / q ( 1 | Q | Q ω ( x ) q d x ) 1 / q C | Q | ω ( Q ) b 1 BMO f L ( ω ) ω ( Q ) | Q | C b 1 BMO f L ( ω ) .

For I 5 (x), we have

I 5 ( x ) = | R n ( b 1 ( y ) ( b 1 ) Q ) ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | C k = 0 2 k + 1 Q 2 k Q | b 1 ( y ) ( b 1 ) Q | | f ( y ) | d δ | x 0 y | n + δ d y C k = 1 d δ ( 2 k 1 d ) n + δ | 2 k Q | ( 1 | 2 k Q | 2 k Q | f ( y ) | p d y ) 1 / p × ( 1 | 2 k Q | 2 k Q | b 1 ( y ) ( b 1 ) Q | p d y ) 1 / p C k = 1 k m 2 k δ b 1 BMO f L ( ω ) C b 1 BMO f L ( ω ) ,

so

1 ω ( Q ) Q | I 5 (x)|ω(x)dxC b 1 BMO f L ( ω ) .

When m>1, set b Q =( ( b 1 ) Q ,, ( b m ) Q ) R n , where ( b j ) Q = | Q | 1 Q b j (y)dy, 1jm, we have

T b ( f ) ( x ) = j = 1 m ( b j ( x ) ( b j ) Q ) R n K ( x , y ) f ( y ) d y + j = 1 m 1 σ C j m ( 1 ) m j ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K ( x , y ) f ( y ) d y + ( 1 ) m R n j = 1 m ( b j ( y ) ( b j ) Q ) K ( x , y ) f ( y ) d y

and

A t Q T b ( f ) ( x ) = j = 1 m ( b j ( x ) ( b j ) Q ) R n K t ( x , y ) f ( y ) d y + j = 1 m 1 σ C j m ( 1 ) m j ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K t ( x , y ) f ( y ) d y + ( 1 ) m R n j = 1 m ( b j ( y ) ( b j ) Q ) K t ( x , y ) f ( y ) d y ,

then

| T b ( f ) ( x ) A t Q T b ( f ) ( x ) | | j = 1 m ( b j ( x ) ( b j ) Q ) R n K ( x , y ) f ( y ) d y | + | j = 1 m 1 σ C j m ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K ( x , y ) f ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) K ( x , y ) f 1 ( y ) d y | + | j = 1 m ( b j ( x ) ( b j ) Q ) R n K t ( x , y ) f ( y ) d y | + | j = 1 m 1 σ C j m ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K t ( x , y ) f ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) K t ( x , y ) f 1 ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | = J 1 ( x ) + J 2 ( x ) + J 3 ( x ) + J 4 ( x ) + J 5 ( x ) + J 6 ( x ) + J 7 ( x ) .

For J 1 (x), same as m=1, for some 1<q<, let 1/ q 1 +1/ q 2 ++1/ q m +1/q=1, 1/p+1/ p =1, by Hölder’s inequality, and the reverse of Hölder’s inequality, we get

1 ω ( Q ) Q | J 1 ( x ) | ω ( x ) d x C ω ( Q ) ( Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) | p ω ( x ) d x ) 1 / p × ( Q | T ( f ) ( x ) | p ω ( x ) d x ) 1 / p C ω ( Q ) ( Q | b 1 ( x ) ( b 1 ) Q | p | b m ( x ) ( b m ) Q | p ω ( x ) d x ) 1 / p × f L ( ω ) ( Q ω ( x ) d x ) 1 / p C ω ( Q ) f L ( ω ) ω ( Q ) 1 / p j = 1 m [ ( Q | b j ( x ) ( b j ) Q | p q j d x ) 1 / q j ( Q ω ( x ) q d x ) 1 / q ] 1 / p C b BMO f L ( ω ) ω ( Q ) 1 / p + 1 / p 1 | Q | 1 / p ( 1 / q 1 + + 1 / q m + 1 / q 1 ) C b BMO f L ( ω ) .

For J 2 (x), by Hölder’s inequality and the reverse of Hölder’s inequality, we have

1 ω ( Q ) Q | J 2 ( x ) | ω ( x ) d x j = 1 m 1 σ C j m C ω ( Q ) ( Q | ( b ( x ) b Q ) σ | p ω ( x ) d x ) 1 / p × ( Q | T ( ( b b Q ) σ c f ) ( x ) | p ω ( x ) d x ) 1 / p C j = 1 m 1 σ C j m ( 1 ω ( Q ) Q | ( b ( x ) b Q ) σ | p ω ( x ) d x ) 1 / p × ( 1 ω ( Q ) Q | T ( ( b b Q ) σ c f ) ( x ) | p ω ( x ) d x ) 1 / p C j = 1 m 1 σ C j m ω ( Q ) 1 / p [ ( Q | ( b ( x ) b Q ) σ | p q d x ) 1 / q ( Q ω ( x ) q d x ) 1 / q ] 1 / p × ω ( Q ) 1 / p ( R n | ( b ( x ) b Q ) σ c f ( x ) | p ω ( x ) χ Q ( x ) d x ) 1 / p C j = 1 m 1 σ C j m ω ( Q ) 1 / p | Q | 1 / p q + 1 / p q 1 / p ω ( Q ) 1 / p b σ BMO × ω ( Q ) 1 / p ( Q | ( b ( x ) b Q ) σ c | p q d x ) 1 / p q ( Q | f ( x ) | p q ω q ( x ) d x ) 1 / p q C j = 1 m 1 σ C j m b σ BMO b σ c BMO ( | Q | ω ( Q ) ) 1 / p × ( 1 | Q | Q ω ( x ) d x ) 1 / p f L ( ω ) C b BMO f L ( ω ) .

For J 3 (x), taking p>1, by the L p (ω)-boundedness of T, we have

1 ω ( Q ) Q | J 3 ( x ) | ω ( x ) d x ( 1 ω ( Q ) R n | T ( ( b 1 ( b 1 ) Q ) ( b m ( b m ) Q ) f 1 ) ( x ) | p ω ( x ) d x ) 1 / p C ω ( Q ) 1 / p ( R n | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) f 1 ( x ) | p ω ( x ) d x ) 1 / p C ω ( Q ) 1 / p | Q | 1 / p q b BMO | Q | 1 / p q ( 1 | Q | Q ω q d x ) 1 / p q f L ( ω ) C b BMO ( | Q | ω ( Q ) ) 1 / p ( ω ( Q ) | Q | ) 1 / p f L ( ω ) C b BMO f L ( ω ) .

For J 4 (x), J 5 (x), and J 6 (x), choose 1<p, q j <, j=1,,m, such that 1/p+1/ q 1 ++1/ q m +1/q, by Lemma 2 and similar to the proofs of J 1 (x), J 2 (x), and J 3 (x), we get

1 ω ( Q ) Q | J 4 ( x ) | ω ( x ) d x C | Q | ω ( Q ) j = 1 m ( 1 | Q | Q | ( b j ( x ) ( b j ) Q ) | q j d x ) 1 / q j × ( 1 | Q | Q | A t Q ( f ) ( x ) | p d x ) 1 / p ( 1 | Q | Q ω ( x ) q d x ) 1 / q C b BMO f L ( ω ) , 1 ω ( Q ) Q | J 5 ( x ) | ω ( x ) d x C | Q | ω ( Q ) j = 1 m 1 σ C j m ( 1 | Q | Q | ( b ( x ) b Q ) σ | q d x ) 1 / q × ( 1 | Q | Q | A t Q ( ( b b Q ) σ c f ) ( x ) | p d x ) 1 / p ( 1 | Q | Q ω ( x ) q d x ) 1 / q C | Q | ω ( Q ) j = 1 m 1 σ C j m b σ BMO b σ c BMO f L ( ω ) ω ( Q ) | Q | C b BMO f L ( ω ) , 1 ω ( Q ) Q | J 6 ( x ) | ω ( x ) d x C | Q | ω ( Q ) ( 1 | Q | Q | A t Q ( ( b 1 ( b 1 ) Q ) ( b m ( b m ) Q ) f 1 ) ( x ) | q d x ) 1 / q × ( 1 | Q | Q ω ( x ) q d x ) 1 / q C b BMO f L ( ω ) .

For J 7 (x), note that |xy|d= t 1 / 2 , taking 1< q j <, j=1,,m such that 1/ q 1 ++1/ q m +1/r=1, then

J 7 ( x ) C k = 0 2 k + 1 Q 2 k Q j = 1 m | ( b j ( y ) ( b j ) Q ) | | f ( y ) | d δ | x 0 y | n + δ d y C k = 1 d δ ( 2 k 1 d ) n + δ | 2 k Q | ( 1 | 2 k Q | 2 k Q | f ( y ) | r d y ) 1 / r × j = 1 m ( 1 | 2 k Q | 2 k Q | b j ( y ) ( b j ) Q | q j d y ) 1 / q j C k = 1 2 k δ f L ( ω ) j = 1 m ( 1 | 2 k Q | 2 k Q | b j ( y ) ( b j ) Q | q j d y ) 1 / q j C k = 1 k m 2 k δ j = 1 m b j BMO f L ( ω ) C b BMO f L ( ω ) ,

so

1 ω ( Q ) Q | J 7 (x)|ω(x)dxC b BMO f L ( ω ) .

This completes the proof of Theorem 1. □

Theorem 2 Let 1<p<, ω A 1 and b =( b 1 ,, b m ) with b j BMO( R n ) for 1jm. Then T b is bounded from B p (ω) to CMO A (ω).

Proof It suffices to prove for f C 0 ( R n ), the following inequality holds:

1 ω ( Q ) Q | T b (f)(x) A t Q T b (f)(x)|ω(x)dxC f B p ( ω )

for any cube Q=Q(0,d) with d>1. Fix a cube Q=Q(0,d) with d>1. Set f 1 =f χ Q , f 2 =f χ ( R n Q ) and b Q =( ( b 1 ) Q ,, ( b m ) Q ) R n , where ( b j ) Q = | Q | 1 Q | b j (y)|dy, 1jm, we have

| T b ( f ) ( x ) A t Q T b ( f ) ( x ) | | j = 1 m ( b j ( x ) ( b j ) Q ) R n K ( x , y ) f ( y ) d y | + | j = 1 m 1 σ C j m ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K ( x , y ) f ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) K ( x , y ) f 1 ( y ) d y | + | j = 1 m ( b j ( x ) ( b j ) Q ) R n K t ( x , y ) f ( y ) d y | + | j = 1 m 1 σ C j m ( b ( x ) ( b ) Q ) σ R n ( b ( y ) ( b ) Q ) σ c K t ( x , y ) f ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) K t ( x , y ) f 1 ( y ) d y | + | R n j = 1 m ( b j ( y ) ( b j ) Q ) ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | = L 1 ( x ) + L 2 ( x ) + L 3 ( x ) + L 4 ( x ) + L 5 ( x ) + L 6 ( x ) + L 7 ( x ) .

For L 1 (x), we have

1 ω ( Q ) Q | L 1 ( x ) | ω ( x ) d x C ω ( Q ) ( Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) | p ω ( x ) d x ) 1 / p × ( Q | T ( f ) ( x ) | p ω ( x ) d x ) 1 / p C ω ( Q ) [ ( Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) | p q d x ) 1 / q ( Q ω ( x ) q d x ) 1 / q ] 1 / p × ( Q | f ( x ) | p ω ( x ) d x ) 1 / p C ω ( Q ) | Q | 1 / p q b BMO | Q | 1 / p q ( ω ( Q ) | Q | ) 1 / p f χ Q L p ( ω ) C b BMO ω ( Q ) 1 / p f χ Q L p ( ω ) C b BMO f B p ( ω ) .

For L 2 (x), taking 1<s, s <, and 1/s+1/ s =1, we have

1 ω ( Q ) Q | L 2 ( x ) | ω ( x ) d x C j = 1 m 1 σ C j m ( 1 ω ( Q ) Q | ( b ( x ) b Q ) σ | s ω ( x ) d x ) 1 / s × ( 1 ω ( Q ) Q | T ( ( b b Q ) σ c f ) ( x ) | s ω ( x ) d x ) 1 / s C j = 1 m 1 σ C j m ω ( Q ) 1 / s [ ( Q | ( b ( x ) b Q ) σ | s q d x ) 1 / q ( Q ω q d x ) 1 / q ] 1 / s × ω ( Q ) 1 / s ( Q | ( b ( x ) b Q ) σ c f ( x ) | s ω ( x ) d x ) 1 / s C j = 1 m 1 σ C j m ω ( Q ) 1 / s | Q | 1 / s q + 1 / s q 1 / s ω ( Q ) 1 / s b σ BMO × ω ( Q ) 1 / s | Q | 1 / r s b σ c BMO ( Q | f ( x ) | p ω ( x ) d x ) 1 / p ( Q ω ( x ) q d x ) ( p s ) / p q s C j = 1 m 1 σ C j m b σ BMO b σ c BMO ω ( Q ) 1 / p f χ Q L p ( ω ) C b BMO f B p ( ω ) .

For L 3 (x), we have

1 ω ( Q ) Q | L 3 ( x ) | ω ( x ) d x C ( 1 ω ( Q ) R n | T ( ( b 1 ( b 1 ) Q ) ( b m ( b m ) Q ) f 1 ) ( x ) | s ω ( x ) d x ) 1 / s C ω ( Q ) 1 / s ( Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) f ( x ) | s ω ( x ) d x ) 1 / s C ω ( Q ) 1 / p b BMO f χ Q L p ( ω ) C b BMO f B p ( ω ) .

For L 4 (x), L 5 (x), and L 6 (x), by Lemma 2, we have

1 ω ( Q ) Q | L 4 ( x ) | ω ( x ) d x C ( 1 ω ( Q ) Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) | s ω ( x ) d x ) 1 / s × ( 1 ω ( Q ) Q | A t Q ( f ) ( x ) | s ω ( x ) d x ) 1 / s C ( 1 ω ( Q ) ) 1 / s [ ( Q | ( b 1 ( x ) ( b 1 ) Q ) ( b m ( x ) ( b m ) Q ) | s q d x ) 1 / q × ( Q ω ( x ) q d x ) 1 / q ] 1 / s f B p ( w ) C ( 1 ω ( Q ) ) 1 / s | Q | 1 / s q b BMO | Q | 1 / s q ( ω ( Q ) | Q | ) 1 / s f B p ( w ) C b BMO f B p ( ω ) ; 1 ω ( Q ) Q | L 5 ( x ) | ω ( x ) d x C j = 1 m 1 σ C j m ( 1 ω ( Q ) Q | ( b ( x ) b Q ) σ | s ω ( x ) d x ) 1 / s × ( 1 ω ( Q ) Q | A t Q ( ( b b Q ) σ c f ) ( x ) | s ω ( x ) d x ) 1 / s C j = 1 m 1 σ C j m ω ( Q ) 1 / s [ ( Q | ( b ( x ) b Q ) σ | s q d x ) 1 / q ( Q ω q d x ) 1 / q ] 1 / s × b σ c BMO f B p ( ω ) C j = 1 m 1 σ C j m ω ( Q ) 1 / s | Q | 1 / s q + 1 / s q 1 / s ω ( Q ) 1 / s b σ BMO b σ c BMO f B p ( ω ) C b σ BMO f B p ( ω ) ; 1 ω ( Q ) Q | L 6 ( x ) | ω ( x ) d x ( 1 ω ( Q ) R n | A t Q ( ( b 1 ( b 1 ) Q ) ( b m ( b m ) Q ) f 1 ) ( x ) | s ω ( x ) d x ) 1 / s C b BMO f B p ( ω ) .

For L 7 (x), note that |xy|d= t 1 / 2 , taking 1<u<p, then

L 7 ( x ) C Q c j = 1 m | b j ( y ) ( b j ) Q | | f ( y ) | d δ | x 0 y | n + δ d y C k = 0 2 k + 1 Q 2 k Q j = 1 m | b j ( y ) ( b j ) Q | | f ( y ) | d δ | x 0 y | n + δ d y C k = 1 d δ ( 2 k 1 d ) n + δ | 2 k Q | ( 1 | 2 k Q | 2 k Q | f ( y ) | u d y ) 1 / u × ( 1 | 2 k Q | 2 k Q j = 1 m | b j ( y ) ( b j ) Q | u d y ) 1 / u C b BMO k = 1 k m 2 k δ ( 1 | 2 k Q | ) 1 / u × [ ( 2 k Q | f ( y ) | p ω ( y ) d y ) u p ( 2 k Q ω ( y ) u p u d y ) p u p ] 1 / u C b BMO k = 1 k m 2 k δ ( 1 | 2 k Q | ) 1 / u f χ 2 k Q L p ( ω ) ( ω ( 2 k Q ) | 2 k Q | ) 1 / p | 2 k Q | ( p u 1 ) 1 p × [ ( 1 | 2 k Q | 2 k Q ω ( y ) d y ) ( 1 | 2 k Q | 2 k Q ω ( y ) 1 p u 1 d y ) p u 1 ] 1 / p C b BMO k = 1 k m 2 k δ ω ( 2 k Q ) 1 / p f χ 2 k Q L p ( ω ) C b BMO f B p ( ω ) ,

so

1 ω ( Q ) Q | L 7 (x)|ω(x)dxC b BMO f B p ( ω ) .

This completes the proof of Theorem 2. □

References

  1. Coifman R, Rochberg R, Weiss G: Factorization theorems for Hardy spaces in several variables. Ann. Math. 1976, 103: 611-635. 10.2307/1970954

    Article  MathSciNet  MATH  Google Scholar 

  2. Harboure E, Segovia C, Torrea JL: Boundedness of commutators of fractional and singular integrals for the extreme values of p . Ill. J. Math. 1997, 41: 676-700.

    MathSciNet  MATH  Google Scholar 

  3. Pérez C, Pradolini G: Sharp weighted endpoint estimates for commutators of singular integral operators. Mich. Math. J. 2001, 49: 23-37. 10.1307/mmj/1008719033

    Article  MATH  Google Scholar 

  4. Deng DG, Yan LX: Commutators of singular integral operators with non-smooth kernels. Acta Math. Sci. 2005, 25: 137-144.

    MathSciNet  MATH  Google Scholar 

  5. Duong XT, McIntosh A: Singular integral operators with non-smooth kernels on irregular domains. Rev. Mat. Iberoam. 1999, 15: 233-265.

    Article  MathSciNet  MATH  Google Scholar 

  6. Garcia-Cuerva J, Rubio de Francia JL North-Holland Mathematics Studies 116. In Weighted Norm Inequalities and Related Topics. North-Holland, Amsterdam; 1985.

    Google Scholar 

  7. Liu LZ: Sharp function boundedness for vector-valued multilinear singular integral operators with non-smooth kernels. J. Contemp. Math. Anal. 2010, 45: 185-196. 10.3103/S1068362310040011

    Article  MathSciNet  MATH  Google Scholar 

  8. Liu LZ: Multilinear singular integral operators on Triebel-Lizorkin and Lebesgue spaces. Bull. Malays. Math. Soc. 2012, 35: 1075-1086.

    MathSciNet  MATH  Google Scholar 

  9. Martell JM: Sharp maximal functions associated with approximations of the identity in spaces of homogeneous type and applications. Stud. Math. 2004, 161: 113-145. 10.4064/sm161-2-2

    Article  MathSciNet  MATH  Google Scholar 

  10. Pérez C, Trujillo-Gonzalez R: Sharp weighted estimates for multilinear commutators. J. Lond. Math. Soc. 2002, 65: 672-692. 10.1112/S0024610702003174

    Article  MathSciNet  MATH  Google Scholar 

  11. Stein EM: Harmonic Analysis: Real Variable Methods, Orthogonality and Oscillatory Integrals. Princeton University Press, Princeton; 1993.

    MATH  Google Scholar 

  12. Torchinsky A Pure and Applied Mathematics 123. In Real Variable Methods in Harmonic Analysis. Academic Press, New York; 1986.

    Google Scholar 

Download references

Acknowledgements

Project was supported by the National Natural Science Foundation of China (No. 11061003).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Mingjun Zhang.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

The authors completed the paper, and read and approved the final manuscript.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0), which permits use, duplication, adaptation, distribution, and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhang, M., Guo, Y. Weighted endpoint estimates for multilinear commutator of singular integral operators with non-smooth kernels. J Inequal Appl 2014, 371 (2014). https://doi.org/10.1186/1029-242X-2014-371

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/1029-242X-2014-371

Keywords