Open Access

A sharp inequality for multilinear singular integral operators with non-smooth kernels

Journal of Inequalities and Applications20132013:439

https://doi.org/10.1186/1029-242X-2013-439

Received: 17 March 2013

Accepted: 12 August 2013

Published: 16 September 2013

Abstract

In this paper, we establish a sharp inequality for some multilinear singular integral operators with non-smooth kernels. As an application, we obtain the weighted L p -norm inequality and L log L -type inequality for the multilinear operators.

MSC:42B20, 42B25.

Keywords

multi-linear operatorsingular integral operator with non-smooth kernelsharp inequalityBMO A p -weight

1 Definitions and results

As the development of singular integral operators and their commutators, multilinear singular integral operators have been well studied (see [16]). In this paper, we study some multilinear operator associated to the singular integral operators with non-smooth kernels as follows.

Definition 1 A family of operators D t , t > 0 , is said to be ‘approximations 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 ) d y
for every f L p ( R n ) with p 1 , 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 a singular integral operator with non-smooth kernel 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 ) d y
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 an 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 ) | d x 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 an associated kernel K t ( x , y ) which satisfies
    | K t ( x , y ) | c 4 t n / 2 if  | x y | c 3 t 1 / 2 ,
     
and
| K ( x , y ) K t ( x , y ) | c 4 t δ / 2 | x y | n δ if  | x y | c 3 t 1 / 2

for some c 3 , c 4 > 0 , δ > 0 .

Let m j be positive integers ( j = 1 , , l ), m 1 + + m l = m , and let b j be functions on R n ( j = 1 , , l ). Set, for 1 j m ,
R m j + 1 ( b j ; x , y ) = b j ( x ) | α | m j 1 α ! D α b j ( y ) ( x y ) α .
The multilinear operator associated to T is defined by
T b ( f ) ( x ) = R n j = 1 l R m j + 1 ( b j ; x , y ) | x y | m K ( x , y ) f ( y ) d y .

Note that when m = 0 , T b is just the multilinear commutator of T and b j (see [7]). However, when m > 0 , T b is a non-trivial generalization of the commutator. It is well known that multilinear operators are of great interest in harmonic analysis and have been widely studied by many authors (see [14]). Hu and Yang (see [8]) proved a variant sharp estimate for the multilinear singular integral operators. In [7], Pérez and Trujillo-Gonzalez proved a sharp estimate for the multilinear commutator when b j Osc exp L r j ( R n ) and noted that Osc exp L r j BMO . The main purpose of this paper is to prove a sharp function inequality for the multilinear singular integral operator with non-smooth kernel when D α b j BMO ( R n ) for all α with | α | = m j . As an application, we obtain an L p ( p > 1 ) norm inequality and an L log L -type inequality for the multilinear operators. In [912], the boundedness of a singular integral operator with non-smooth kernel is obtained. In [13], the boundedness of the commutator associated to the singular integral operator with non-smooth kernel is obtained. Our works are motivated by these papers.

First, let us introduce some notations. Throughout this paper, Q denotes a cube of R n with sides parallel to the axes. For any locally integrable function f, the sharp function of f is defined by
f # ( x ) = sup Q x 1 | Q | Q | f ( y ) f Q | d y ,
where, and in what follows, f Q = | Q | 1 Q f ( x ) d x . It is well known that (see [14, 15])
f # ( x ) sup Q x inf c C 1 | Q | Q | f ( y ) c | d y .
We say that f belongs to BMO ( R n ) if f # belongs to L ( R n ) and f BMO = f # L . Let M be a Hardy-Littlewood maximal operator defined by
M ( f ) ( x ) = sup Q x 1 | Q | Q | f ( y ) | d y .
For k N , we denote by M k the operator M iterated k times, i.e., M 1 ( f ) ( x ) = M ( f ) ( x ) and
M k ( f ) ( x ) = M ( M k 1 ( f ) ) ( x ) when  k 2 .
The sharp maximal function M A ( f ) associated with the ‘approximations to the identity’ { A t , t > 0 } is defined by
M A # ( f ) ( x ) = sup Q x 1 | Q | Q | f ( y ) A t Q ( f ) ( y ) | d y ,
where t Q = l ( Q ) 2 and l ( Q ) denotes the side length of Q. For 0 < r < , we denote M A # ( f ) r by
M A # ( f ) r = [ M A # ( | f | r ) ] 1 / r .
Let Φ be a Young function and Φ ˜ be the complementary associated to Φ. For a function f, we denote the Φ-average by
f Φ , Q = inf { λ > 0 : 1 | Q | Q Φ ( | f ( y ) | λ ) d y 1 }
and the maximal function associated to Φ by
M Φ ( f ) ( x ) = sup Q x f Φ , Q .
The Young functions used in this paper are Φ ( t ) = t ( 1 + log t ) r and Φ ˜ ( t ) = exp ( t 1 / r ) , the corresponding average and maximal functions are denoted by L ( log L ) r , Q , M L ( log L ) r and exp L 1 / r , Q , M exp L 1 / r . Following [11, 12, 16], we know the generalized Hölder inequality
1 | Q | Q | f ( y ) g ( y ) | d y f Φ , Q g Φ ˜ , Q
and the following inequality, for r , r j 1 , j = 1 , , l with 1 / r = 1 / r 1 + + 1 / r l , and any x R n , b BMO ( R n ) ,
f L ( log L ) 1 / r , Q M L ( log L ) 1 / r ( f ) C M L ( log L ) l ( f ) C M l + 1 ( f ) , b b Q exp L r , Q C b BMO , | b 2 k + 1 Q b 2 Q | C k b BMO .

We denote the Muckenhoupt weights by A p for 1 p < (see [14]).

We shall prove the following theorems.

Theorem 1 If T is a singular integral operator with non-smooth kernel as given in Definition  2, let D α b j BMO ( R n ) for all α with | α | = m j and j = 1 , , l . Then there exists a constant C > 0 such that for any f C 0 ( R n ) , 0 < r < 1 and x ˜ R n ,
M A # ( T b ( f ) ) r ( x ˜ ) C j = 1 l ( | α j | = m j D α j b j BMO ) M l + 1 ( f ) ( x ˜ ) .
Theorem 2 If T is a singular integral operator with non-smooth kernel as given in Definition  2, let D α b j BMO ( R n ) for all α with | α | = m j and j = 1 , , l . Then T b is bounded on L p ( w ) for any 1 < p < and w A p , that is,
T b ( f ) L p ( w ) C j = 1 l ( | α j | = m j D α j b j BMO ) f L p ( w ) .
Theorem 3 If T is a singular integral operator with non-smooth kernel as given in Definition  2, let w A 1 , D α b j BMO ( R n ) for all α with | α | = m j and j = 1 , , l . Then there exists a constant C > 0 such that for all λ > 0 ,
w ( { x R n : | T b ( f ) ( x ) | > λ } ) C R n | f ( x ) | λ [ 1 + log + ( | f ( x ) | λ ) ] l w ( x ) d x .

2 Proof of the theorem

To prove the theorems, we need the following lemma.

Lemma 1 (see [1])

Let b be a function on R n and D α b L q ( R n ) for all α with | α | = m and some q > n . Then
| R m ( b ; x , y ) | C | x y | m | α | = m ( 1 | Q ˜ ( x , y ) | Q ˜ ( x , y ) | D α b ( z ) | q d z ) 1 / q ,

where Q ˜ is the cube centered at x and having side length 5 n | x y | .

Lemma 2 ([[14], p.485])

Let 0 < p < q < and for any function f 0 , we define that for 1 / r = 1 / p 1 / q ,
f W L q = sup λ > 0 λ | { x R n : f ( x ) > λ } | 1 / q , N p , q ( f ) = sup E f χ E L p / χ E L r ,
where the sup is taken for all measurable sets E with 0 < | E | < . Then
f W L q N p , q ( f ) ( q / ( q p ) ) 1 / p f W L q .

Lemma 3 (see [17])

Let r j 1 for j = 1 , , l , we denote that 1 / r = 1 / r 1 + + 1 / r l . Then
1 | Q | Q | f 1 ( x ) f l ( x ) g ( x ) | d x f exp L r 1 , Q f exp L r l , Q g L ( log L ) 1 / r , Q .

Lemma 4 ([9, 10])

Let T be a singular integral operator with non-smooth kernel as given in Definition  2. Then T is bounded on L p ( R n ) for every 1 < p < and bounded from L 1 ( R n ) to W L 1 ( R n ) .

Lemma 5 (see [9, 12])

For any γ > 0 , there exists a constant C > 0 independent of γ such that
| { x R n : M ( f ) ( x ) > D λ , M A # ( f ) ( x ) γ λ } | C γ | { x R n : M ( f ) ( x ) > λ } |
for λ > 0 , where D is a fixed constant which only depends on n. Thus
M ( f ) L p C M A # ( f ) L p

for every f L p ( R n ) , 1 < p < .

Lemma 6 Let { A t , t > 0 } be an ‘approximation to the identity’ and b BMO ( R n ) . Then, for every f L p ( R n ) , p > 1 and x ˜ R n ,
sup Q x ˜ 1 | Q | Q | A t Q ( ( b b Q ) f ) ( x ) | d x C b BMO M 2 ( f ) ( x ˜ ) ,

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

Proof We write, for any cube Q with x ˜ Q ,
1 | Q | Q | A t Q ( ( b b Q ) f ) ( x ) | d x 1 | Q | Q R n h t Q ( x , y ) | ( b ( y ) b Q ) f ( y ) | d y d x 1 | Q | Q 2 Q h t Q ( x , y ) | ( b ( y ) b Q ) f ( y ) | d y d x + k = 1 1 | Q | Q 2 k + 1 Q 2 k Q h t Q ( x , y ) | ( b ( y ) b Q ) f ( y ) | d y d x = I 1 + I 2 .
We have, by the generalized Hölder inequality,
I 1 C 1 | Q | | 2 Q | Q 2 Q | ( b ( y ) b Q ) f ( y ) | d y d x C b b Q exp L , 2 Q f L ( log L ) , 2 Q C b BMO M 2 ( f ) ( x ˜ ) .
For I 2 , notice for x Q and y 2 k + 1 Q 2 k Q , then | x y | 2 k 1 t Q and h t Q ( x , y ) C s ( 2 2 ( k 1 ) ) | Q | , then
I 2 C k = 1 s ( 2 2 ( k 1 ) ) 1 | Q | 2 Q 2 k + 1 Q | ( b ( y ) b Q ) f ( y ) | d y d x C k = 1 2 k n s ( 2 2 ( k 1 ) ) 1 | 2 k + 1 Q | 2 k + 1 Q | ( b ( y ) b Q ) f ( y ) | d y C k = 1 2 k n s ( 2 2 ( k 1 ) ) b b Q exp L , 2 k + 1 Q f L ( log L ) , 2 k + 1 Q C k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) b BMO M 2 ( f ) ( x ˜ ) C b BMO M 2 ( f ) ( x ˜ ) ,
where the last inequality follows from
k = 1 2 ( k 1 ) n s ( 2 2 ( k 1 ) ) C k = 1 2 ( k 1 ) ε <

for some ϵ > 0 . This completes the proof. □

Proof of Theorem 1 It suffices to prove for f C 0 ( R n ) and some constant C 0 that the following inequality holds:
( 1 | Q | Q | | T b ( f ) ( x ) | r | A t Q T b ( f ) ( x ) | r | d x ) 1 / r C j = 1 l ( | α j | = m j D α j b j BMO ) M l + 1 ( f ) ( x ) .
Without loss of generality, we may assume l = 2 . Fix a cube Q = Q ( x 0 , d ) and x ˜ Q . Let Q ˜ = 5 n Q and b ˜ j ( x ) = b j ( x ) | α | = m j 1 α ! ( D α b j ) Q ˜ x α , then R m j ( b j ; x , y ) = R m j ( b ˜ j ; x , y ) and D α b ˜ j = D α b j ( D α b j ) Q ˜ for | α | = m j . We write, for f = f χ Q ˜ + f χ R n Q ˜ = f 1 + f 2 ,
T b ( f ) ( x ) = R n j = 1 2 R m j + 1 ( b ˜ j ; x , y ) | x y | m K ( x , y ) f ( y ) d y = R n j = 1 2 R m j ( b ˜ j ; x , y ) | x y | m K ( x , y ) f 1 ( y ) d y | α 1 | = m 1 1 α 1 ! R n R m 2 ( b ˜ 2 ; x , y ) ( x y ) α 1 D α 1 b ˜ 1 ( y ) | x y | m K ( x , y ) f 1 ( y ) d y | α 2 | = m 2 1 α 2 ! R n R m 1 ( b ˜ 1 ; x , y ) ( x y ) α 2 D α 2 b ˜ 2 ( y ) | x y | m K ( x , y ) f 1 ( y ) d y + | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! R n ( x y ) α 1 + α 2 D α 1 b ˜ 1 ( y ) D α 2 b ˜ 2 ( y ) | x y | m K ( x , y ) f 1 ( y ) d y + R n j = 1 2 R m j + 1 ( b ˜ j ; x , y ) | x y | m K ( x , y ) f 2 ( y ) d y = T ( j = 1 2 R m j ( b ˜ j ; x , ) | x | m f 1 ) T ( | α 1 | = m 1 1 α 1 ! R m 2 ( b ˜ 2 ; x , ) ( x ) α 1 D α 1 b ˜ 1 | x | m f 1 ) T ( | α 2 | = m 2 1 α 2 ! R m 1 ( b ˜ 1 ; x , ) ( x ) α 2 D α 2 b ˜ 2 | x | m f 1 ) + T ( | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! ( x ) α 1 + α 2 D α 1 b ˜ 1 D α 2 b ˜ 2 | x | m f 1 ) + T ( j = 1 2 R m j + 1 ( b ˜ j ; x , ) | x | m f 2 )
and
A t Q T b ( f ) ( x ) = R n j = 1 2 R m j ( b ˜ j ; x , y ) | x y | m K t ( x , y ) f 1 ( y ) d y | α 1 | = m 1 1 α 1 ! R n R m 2 ( b ˜ 2 ; x , y ) ( x y ) α 1 D α 1 b ˜ 1 ( y ) | x y | m K t ( x , y ) f 1 ( y ) d y | α 2 | = m 2 1 α 2 ! R n R m 1 ( b ˜ 1 ; x , y ) ( x y ) α 2 D α 2 b ˜ 2 ( y ) | x y | m K t ( x , y ) f 1 ( y ) d y + | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! R n ( x y ) α 1 + α 2 D α 1 b ˜ 1 ( y ) D α 2 b ˜ 2 ( y ) | x y | m K t ( x , y ) f 1 ( y ) d y + R n j = 1 2 R m j + 1 ( b ˜ j ; x , y ) | x y | m K t ( x , y ) f 2 ( y ) d y = A t Q T ( j = 1 2 R m j ( b ˜ j ; x , ) | x | m f 1 ) A t Q T ( | α 1 | = m 1 1 α 1 ! R m 2 ( b ˜ 2 ; x , ) ( x ) α 1 D α 1 b ˜ 1 | x | m f 1 ) A t Q T ( | α 2 | = m 2 1 α 2 ! R m 1 ( b ˜ 1 ; x , ) ( x ) α 2 D α 2 b ˜ 2 | x | m f 1 ) + A t Q T ( | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! ( x ) α 1 + α 2 D α 1 b ˜ 1 D α 2 b ˜ 2 | x | m f 1 ) + A t Q T ( j = 1 2 R m j + 1 ( b ˜ j ; x , ) | x | m f 2 ) ,
then
[ 1 | Q | Q | | T b ( f ) ( x ) | r | A t Q T b ( f ) ( x ) | r | d x ] 1 / r [ 1 | Q | Q | T b ( f ) ( x ) A t Q T b ( f ) ( x ) | r d x ] 1 / r [ C | Q | Q | T ( j = 1 2 R m j ( b ˜ j ; x , ) | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | T ( | α 1 | = m 1 R m 2 ( b ˜ 2 ; x , ) ( x ) α 1 D α 1 b ˜ 1 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | T ( | α 2 | = m 2 R m 1 ( b ˜ 1 ; x , ) ( x ) α 2 D α 2 b ˜ 2 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | T ( | α 1 | = m 1 , | α 2 | = m 2 Q ( x ) α 1 + α 2 D α 1 b ˜ 1 D α 2 b ˜ 2 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | A t Q T ( j = 1 2 R m j ( b ˜ j ; x , ) | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | A t Q T ( | α 1 | = m 1 1 α 1 ! R m 2 ( b ˜ 2 ; x , ) ( x ) α 1 D α 1 b ˜ 1 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | A t Q T ( | α 2 | = m 2 1 α 2 ! R m 1 ( b ˜ 1 ; x , ) ( x ) α 2 D α 2 b ˜ 2 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | A t Q T ( | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! ( x ) α 1 + α 2 D α 1 b ˜ 1 D α 2 b ˜ 2 | x | m f 1 ) | r d x ] 1 / r + [ C | Q | Q | ( T A t Q T ) ( j = 1 2 R m j + 1 ( b ˜ j ; x , ) | x | m f 2 ) | r d x ] 1 / r : = I 1 + I 2 + I 3 + I 4 + I 5 + I 6 + I 7 + I 8 + I 9 .
Now, let us estimate I 1 , I 2 , I 3 , I 4 , I 5 , I 6 , I 7 , I 8 and I 9 , respectively. First, for x Q and y Q ˜ , by Lemma 1, we get
R m ( b ˜ j ; x , y ) C | x y | m | α j | = m D α j b j BMO ,
by Lemma 2 and the weak type ( 1 , 1 ) of T (Lemma 4), we obtain
I 1 C j = 1 2 ( | α j | = m j D α j b j BMO ) ( 1 | Q | R n | T ( f 1 ) ( x ) | r d x ) 1 / r C j = 1 2 ( | α j | = m j D α j b j BMO ) | Q | 1 T ( f 1 ) χ Q L r | Q | 1 / r 1 C j = 1 2 ( | α j | = m j D α j b j BMO ) | Q | 1 T ( f 1 ) W L 1 C j = 1 2 ( | α j | = m j D α j b j BMO ) | Q ˜ | 1 f 1 L 1 C j = 1 2 ( | α j | = m j D α j b j BMO ) M ( f ) ( x ˜ ) .
For I 2 , we get, by Lemma 2 and the generalized Hölder inequality,
I 2 C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 ( 1 | Q | R n | T ( D α 1 b ˜ 1 f 1 ) ( x ) | r d x ) 1 / r C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 | Q | 1 T ( D α 1 b ˜ 1 f 1 ) χ Q L r | Q | 1 / r 1 C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 | Q | 1 T ( D α 1 b ˜ 1 f 1 ) W L 1 C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 | Q ˜ | 1 D α 1 b ˜ 1 f 1 L 1 C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 D α 1 b 1 ( D α b 1 ) Q ˜ exp L , Q ˜ f L ( log L ) , Q ˜ C j = 1 2 ( | α j | = m j D α j b j BMO ) M 2 ( f ) ( x ˜ ) .
For I 3 , similar to the proof of I 2 , we get
I 3 C j = 1 2 ( | α | = m j D α b j BMO ) M 2 ( f ) ( x ˜ ) .
Similarly, for I 4 , taking r , r 1 , r 2 1 such that 1 / r = 1 / r 1 + 1 / r 2 , we obtain, by Lemma 3 and the generalized Hölder inequality,
I 4 C | α 1 | = m 1 , | α 2 | = m 2 ( 1 | Q | R n | T ( D α 1 b ˜ 1 D α 2 b ˜ 2 f 1 ) ( x ) | r d x ) 1 / r C | α 1 | = m 1 , | α 2 | = m 2 | Q | 1 T ( D α 1 b ˜ 1 D α 2 b ˜ 2 f 1 ) χ Q L r | Q | 1 / r 1 C | α 1 | = m 1 , | α 2 | = m 2 | Q | 1 T ( D α 1 b ˜ 1 D α 2 b ˜ 2 f 1 ) W L 1 C | α 1 | = m 1 , | α 2 | = m 2 | Q | 1 D α 1 b ˜ 1 D α 2 b ˜ 2 f 1 L 1 C | α 1 | = m 1 , | α 2 | = m 2 j = 1 2 D α j b j ( D α j b j ) Q ˜ exp L r j , Q ˜ f L ( log L ) 1 / r , Q ˜ C j = 1 2 ( | α | = m j D α b j BMO ) M 3 ( f ) ( x ˜ ) .
For I 5 , I 6 , I 7 and I 8 , by Lemma 6, we get
I 5 + I 6 + I 7 + I 8 C j = 1 2 ( | α j | = m j D α j b j BMO ) 1 | Q | Q | A t Q T ( f 1 ) ( x ) | d x + C | α 2 | = m 2 D α 2 b 2 BMO | α 1 | = m 1 1 | Q | Q | A t Q T ( D α 1 b ˜ 1 f 1 ) ( x ) | d x + C | α 1 | = m 1 D α 1 b 1 BMO | α 2 | = m 2 1 | Q | Q | A t Q T ( D α 2 b ˜ 2 f 1 ) ( x ) | d x + C | α 1 | = m 1 , | α 2 | = m 2 1 | Q | Q | A t Q T ( D α 1 b ˜ 1 D α 2 b ˜ 2 f 1 ) ( x ) | d x C j = 1 2 ( | α | = m j D α b j BMO ) M 3 ( f ) ( x ˜ ) .
For I 9 , we write
( T A t Q T ) ( j = 1 2 R m j + 1 ( b ˜ j ; x , ) | x | m f 2 ) = R n j = 1 2 R m j + 1 ( b ˜ j ; x , y ) | x y | m ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y = R n j = 1 2 R m j ( b ˜ j ; x , y ) | x y | m ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | α 1 | = m 1 1 α 1 ! R n D α 1 b ˜ 1 ( y ) ( x y ) α 1 R m 2 ( b ˜ 2 ; x , y ) | x y | m ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y | α 2 | = m 2 1 α 2 ! R n D α 2 b ˜ 2 ( y ) ( x y ) α 2 R m 1 ( b ˜ 1 ; x , y ) | x y | m ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y + | α 1 | = m 1 , | α 2 | = m 2 1 α 1 ! α 2 ! R n D α 1 b ˜ 1 ( y ) D α 2 b ˜ 2 ( y ) ( x y ) α 1 + α 2 | x y | m ( K ( x , y ) K t ( x , y ) ) f 2 ( y ) d y = I 9 ( 1 ) + I 9 ( 2 ) + I 9 ( 3 ) + I 9 ( 4 ) .
By Lemma 1 and the following inequality (see [15])
| b Q 1 b Q 2 | C log ( | Q 2 | / | Q 1 | ) b BMO for  Q 1 Q 2 ,
we know that for x Q and y 2 k + 1 Q ˜ 2 k Q ˜ ,
| R m ( b ˜ ; x , y ) | C | x y | m | α | = m ( D α b BMO + | ( D α b ) Q ˜ ( x , y ) ( D α b ) Q ˜ | ) C k | x y | m | α | = m D α b BMO .
Note that | x y | d = t 1 / 2 and | x y | | x 0 y | for x Q and y R n Q ˜ . By the conditions on K and K t , we obtain
| I 9 ( 1 ) | = k = 0 2 k + 1 Q ˜ 2 k Q ˜ j = 1 2 | R m j ( b ˜ j ; x , y ) | | x y | m | K ( x , y ) K t ( x , y ) | | f ( y ) | d y C j = 1 2 ( | α | = m j D α b j BMO ) k = 0 2 k + 1 Q ˜ 2 k Q ˜ k 2 d δ | x 0 y | n + δ | f ( y ) | d y C j = 1 2 ( | α | = m j D α b j BMO ) k = 1 k 2 2 δ k 1 | 2 k Q ˜ | 2 k Q ˜ | f ( y ) | d y C j = 1 2 ( | α | = m j D α b j BMO ) M ( f ) ( x ˜ ) .
For I 9 ( 2 ) , we get, by the generalized Hölder inequality,
| I 9 ( 2 ) | C ( | α 2 | = m 2 D α 2 b 2 BMO ) | α 1 | = m 1 k = 0 2 k + 1 Q ˜ 2 k Q ˜ k d δ | x 0 y | n + δ | D α 1 b ˜ 1 ( y ) | | f ( y ) | d y C ( | α 2 | = m 2 D α 2 b 2 BMO ) × | α 1 | = m 1 k = 1 k 2 δ k D α 1 b 1 ( D α 1 b 1 ) Q ˜ exp L , 2 k Q ˜ f L ( log L ) , 2 k Q ˜ C j = 1 2 ( | α | = m j D α b j BMO ) M 2 ( f ) ( x ˜ ) .
Similarly,
| I 9 ( 3 ) | C j = 1 2 ( | α | = m j D α b j BMO ) M 2 ( f ) ( x ˜ ) .
For I 9 ( 4 ) , taking r , r 1 , r 2 1 such that 1 / r = 1 / r 1 + 1 / r 2 , by Lemma 3 and the generalized Hölder inequality, we get
| I 9 ( 4 ) | C | α 1 | = m 1 , | α 2 | = m 2 k = 0 2 k + 1 Q ˜ 2 k Q ˜ d δ | x 0 y | n + δ | D α 1 b ˜ 1 ( y ) | | D α 2 b ˜ 2 ( y ) | | f ( y ) | d y C | α 1 | = m 1 , | α 2 | = m 2 k = 1 j = 1 2 D α j b j ( D α j b j ) Q ˜ exp L r j , 2 k Q ˜ f L ( log L ) 1 / r , 2 k Q ˜ C j = 1 2 ( | α | = m j D α b j BMO ) M 3 ( f ) ( x ˜ ) .
Thus
| I 5 | C j = 1 2 ( | α | = m j D α b j BMO ) M 3 ( f ) ( x ˜ ) .

This completes the proof of Theorem 1. □

By Theorem 1 and the L p ( w ) -boundedness of M l + 1 , we may obtain the conclusions of Theorem 2. By Theorem 1 and [16, 17], we may obtain the conclusions of Theorem 3.

3 Applications

In this section we shall apply Theorems 1, 2 and 3 of the paper to the holomorphic functional calculus of linear elliptic operators. First, we review some definitions regarding the holomorphic functional calculus (see [9]). Given 0 θ < π , define
S θ = { z C : | arg ( z ) | θ } { 0 }
and its interior by S θ 0 . Set S ˜ θ = S θ { 0 } . A closed operator L on some Banach space E is said to be of type θ if its spectrum σ ( L ) S θ and if for every ν ( θ , π ] , there exists a constant C ν such that
| η | ( η I L ) 1 C ν , η S ˜ θ .
For ν ( 0 , π ] , let
H ( S μ 0 ) = { f : S θ 0 C : f  is holomorphic and  f L < } ,
where f L = sup { | f ( z ) | : z S μ 0 } . Set
Ψ ( S μ 0 ) = { g H ( S μ 0 ) : s > 0 , c > 0  such that  | g ( z ) | c | z | s 1 + | z | 2 s } .
If L is of type θ and g H ( S μ 0 ) , we define g ( L ) L ( E ) by
g ( L ) = ( 2 π i ) 1 Γ ( η I L ) 1 g ( η ) d η ,
where Γ is the contour { ξ = r e ± i ϕ : r 0 } parameterized clockwise around S θ with θ < ϕ < μ . If, in addition, L is one-to-one and has a dense range, then, for f H ( S μ 0 ) ,
f ( L ) = [ h ( L ) ] 1 ( f h ) ( L ) ,
where h ( z ) = z ( 1 + z ) 2 . L is said to have a bounded holomorphic functional calculus on the sector S μ if
g ( L ) N g L

for some N > 0 and for all g H ( S μ 0 ) .

Now, let L be a linear operator on L 2 ( R n ) with θ < π / 2 so that ( L ) generates a holomorphic semigroup e z L , 0 | arg ( z ) | < π / 2 θ . Applying Theorem 6 of [9], we get the following.

Theorem 4 Assume the following conditions are satisfied:
  1. (i)
    The holomorphic semigroup e z L , 0 | arg ( z ) | < π / 2 θ is represented by the kernels a z ( x , y ) which satisfy, for all ν > θ , an upper bound
    | a z ( x , y ) | c ν h | z | ( x , y )
     
for x , y R n , and 0 | arg ( z ) | < π / 2 θ , where h t ( x , y ) = C t n / 2 s ( | x y | 2 / t ) and s is a positive, bounded and decreasing function satisfying
lim r r n + ϵ s ( r 2 ) = 0 .
  1. (ii)
    The operator L has a bounded holomorphic functional calculus in L 2 ( R n ) , that is, for all ν > θ and g H ( S μ 0 ) , the operator g ( L ) satisfies
    g ( L ) ( f ) L 2 c ν g L f L 2 .
     
Then, for D α b j BMO ( R n ) for all α with | α | = m j and j = 1 , , l , the multilinear operator g ( L ) b associated to g ( L ) and b j satisfies:
  1. (a)
    For 0 < r < 1 and x ˜ R n ,
    M A # ( g ( L ) b ( f ) ) r ( x ˜ ) C j = 1 l ( | α j | = m j D α j b j BMO ) M l + 1 ( f ) ( x ˜ ) ;
     
  2. (b)
    g ( L ) b is bounded on L p ( w ) for any 1 < p < and w A p , that is,
    g ( L ) b ( f ) L p ( w ) C j = 1 l ( | α j | = m j D α j b j BMO ) f L p ( w ) ;
     
  3. (c)
    There exists a constant C > 0 such that for all λ > 0 and w A 1 ,
    w ( { x R n : | g ( L ) b ( f ) ( x ) | > λ } ) C R n | f ( x ) | λ [ 1 + log + ( | f ( x ) | λ ) ] l w ( x ) d x .
     

Declarations

Acknowledgements

Project supported by Hunan Provincial Natural Science Foundation of China (12JJ6003) and Scientific Research Fund of Hunan Provincial Education Departments (12K017).

Authors’ Affiliations

(1)
College of Mathematics and Econometrics, Hunan University

References

  1. Cohen J:A sharp estimate for a multilinear singular integral on R n . Indiana Univ. Math. J. 1981, 30: 693–702. 10.1512/iumj.1981.30.30053MathSciNetView ArticleGoogle Scholar
  2. Cohen J, Gosselin J:On multilinear singular integral operators on R n . Stud. Math. 1982, 72: 199–223.MathSciNetGoogle Scholar
  3. Cohen J, Gosselin J: A BMO estimate for multilinear singular integral operators. Ill. J. Math. 1986, 30: 445–465.MathSciNetGoogle Scholar
  4. Coifman R, Meyer Y Cambridge Studies in Advanced Math. 48. In Wavelets, Calderón-Zygmund and Multilinear Operators. Cambridge University Press, Cambridge; 1997.Google Scholar
  5. Coifman RR, Rochberg R, Weiss G: Factorization theorems for Hardy spaces in several variables. Ann. Math. 1976, 103: 611–635. 10.2307/1970954MathSciNetView ArticleGoogle Scholar
  6. Ding Y, Lu SZ: Weighted boundedness for a class rough multilinear operators. Acta Math. Sin. 2001, 17: 517–526.MathSciNetView ArticleGoogle Scholar
  7. Pérez C, Trujillo-Gonzalez R: Sharp weighted estimates for multilinear commutators. J. Lond. Math. Soc. 2002, 65: 672–692. 10.1112/S0024610702003174View ArticleGoogle Scholar
  8. Hu GE, Yang DC: A variant sharp estimate for multilinear singular integral operators. Stud. Math. 2000, 141: 25–42.MathSciNetGoogle Scholar
  9. Duong XT, McIntosh A: Singular integral operators with non-smooth kernels on irregular domains. Rev. Mat. Iberoam. 1999, 15: 233–265.MathSciNetView ArticleGoogle Scholar
  10. 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/S1068362310040011MathSciNetView ArticleGoogle Scholar
  11. Liu LZ: Multilinear singular integral operators on Triebel-Lizorkin and Lebesgue spaces. Bull. Malays. Math. Soc. 2012, 35: 1075–1086.Google Scholar
  12. 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/sm161-2-2MathSciNetView ArticleGoogle Scholar
  13. Deng DG, Yan LX: Commutators of singular integral operators with non-smooth kernels. Acta Math. Sci. 2005, 25: 137–144.MathSciNetGoogle Scholar
  14. 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
  15. Stein EM: Harmonic Analysis: Real Variable Methods, Orthogonality and Oscillatory Integrals. Princeton University Press, Princeton; 1993.Google Scholar
  16. Pérez C: Endpoint estimate for commutators of singular integral operators. J. Funct. Anal. 1995, 128: 163–185. 10.1006/jfan.1995.1027MathSciNetView ArticleGoogle Scholar
  17. Pérez C, Pradolini G: Sharp weighted endpoint estimates for commutators of singular integral operators. Mich. Math. J. 2001, 49: 23–37. 10.1307/mmj/1008719033View ArticleGoogle Scholar

Copyright

© Gu and Cai; licensee Springer. 2013

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.