Weighted CBMO estimates for commutators of matrix Hausdorff operator on the Heisenberg group

Amna Ajaib; Amjad Hussain

doi:10.1515/math-2020-0175

Open Access Published by De Gruyter Open Access June 4, 2020

Weighted CBMO estimates for commutators of matrix Hausdorff operator on the Heisenberg group

Amna Ajaib and Amjad Hussain

From the journal Open Mathematics

https://doi.org/10.1515/math-2020-0175

Abstract

In this article, we study the commutators of Hausdorff operators and establish their boundedness on the weighted Herz spaces in the setting of the Heisenberg group.

Keywords: Hausdorff operator; commutators; Herz space; CBMO function; Heisenberg group; weights

MSC 2010: 42B35; 42B30; 46E30; 22E25

1 Introduction

The matrix Hausdorff operators defined on n-dimensional Euclidean space ℝ n was first reported in [1], in the following form:

(1.1) H Φ , A f ( x ) = ∫ ℝ n Φ ( y ) f ( x A ( y ) ) d y .

Taking into consideration the duality of the Hardy space H ¹ and bounded mean oscillation (BMO) space, Lerner and Liflyand in [1] have shown that H Φ , A is bounded on Hardy spaces. Subsequently, similar boundedness of H Φ , A was reconsidered in [2] using atomic decomposition of Hardy spaces. The above cited publications are important as their results are the first attempts to study the high-dimensional Hausdorff operators on H 1 ( ℝ n ) . Recently, Liflyand and Miyachi [3] extended these results on H p ( ℝ n ) spaces with 0 < p < 1 .

In 2012, Chen et al. [4] modified the form of (1.1) by replacing the kernel function Φ ( y ) with Φ ( y ) / | y | n :

(1.2) H Φ , A f ( x ) = ∫ ℝ n Φ ( y ) | y | n f ( x A ( y ) ) d y .

As a subcase, when A ( y ) = diag [ 1 / | y | , 1 / | y | , … , 1 / | y | ] , they give another definition of the n-dimensional Hausdorff operator:

(1.3) H Φ f ( x ) = ∫ ℝ n Φ ( y ) | y | n f ( x | y | ) d y .

Their results include the boundedness of Hausdorff operators on Hardy spaces, local Hardy spaces, Herz and Herz-type Hardy spaces with a conclusion that these operators have better performance on Herz-type Hardy spaces than their performance on Hardy spaces. In the same year, with different co-authors, Chen et al. [5] extended the problem on the boundedness of H Φ , A to the product of Hardy-type spaces. The boundedness results regarding Hausdorff operators on H 1 ( ℝ n ) were improved in [6]. The continuity of (1.2) on Morrey spaces, Hardy-Morrey spaces, Block spaces and rectangularly defined spaces has also been discussed in [7], [8], [9] and [10], respectively. Similarly, some results regarding the boundedness of H Φ can be found in [11,12,13].

In the same way, the study of commutators to integral operators is important as it has many applications in the theory of partial differential equations and in characterizing function spaces (see, for instance, [14,15,16]). An attempt has been made in [17] to discuss the boundedness of commutators of H Φ , A , defined by:

(1.4) H Φ , A b f ( x ) = ∫ ℝ n Φ ( y ) | y | n ( b ( x ) − b ( x A ( y ) ) ) f ( x A ( y ) ) d y ,

on function spaces when the symbol function b is either from the Lipschitz space or central BMO space. However, when the matrix A ( y ) is diagonal, we get the commutators of H Φ which were studied in [18,19,20]. For detailed history and other developments regarding Hausdorff operators, we refer the interested readers to the review articles [21,22].

Besides the Euclidean space ℝ n , the matrix Hausdorff operator can be defined on the p-adic linear space ℚ p n , which is a locally compact commutative group under addition (see, for instance, [23,24]), and on the Heisenberg group ℍ n [25,26,27]. Since, we are mainly concerned with the study of the commutators of Hausdorff operators defined on the Heisenberg group ℍ n , therefore, it is mandatory to introduce this group briefly and the definition of matrix Hausdorff operators on it first.

With underlying manifold ℝ 2 n × ℝ , the Heisenberg group ℍ n is the Lie group under the law of non-commutative multiplication

x ⋅ y = ( x 1 , x 2 , … , x 2 n + 1 ) ⋅ ( y 1 , y 2 , … , y 2 n + 1 ) = ( x 1 + y 1 , … , x 2 n + y 2 n , x 2 n + 1 + y 2 n + 1 + 2 ∑ j = 1 n ( y j x n + j − x j y n + j ) ) .

The above definition suggests that for x ∈ ℍ n , we have x ⋅ 0 = x and x ⋅ − x = 0 . Therefore, the identity and inverse elements of ℍ n are same as that of ℝ 2 n + 1 Euclidean space. The basis for the corresponding Lie algebra is formed by the vector fields

X j = ∂ ∂ x j + 2 x n + j ∂ ∂ x 2 n + 1 , 1 ≤ j ≤ n , X n + j = ∂ ∂ x n + j − 2 x j ∂ ∂ x 2 n + 1 , 1 ≤ j ≤ n , X 2 n + 1 = ∂ ∂ x 2 n + 1 .

The only non-vanishing commutator relations satisfied by these vector fields are

[ X j , X n + j ] = − 4 X 2 n + 1 , 1 ≤ j ≤ n .

The dilation, on the Heisenberg group ℍ n , is defined as

δ r ( x 1 , x 2 , … , x 2 n , x 2 n + 1 ) = ( r x 1 , r x 2 , r x 2 n , r 2 x 2 n + 1 ) , r > 0 .

Also, the Haar measure on ℍ n coincides with the usual Lebesgue measure on ℝ 2 n × ℝ 1 . Thus, for any measurable set E ⊂ ℍ n , we denote its measure by | E | . Moreover, it is easy to see that

| δ r ( E ) | = r Q | E | , d ( δ r x ) = r Q d x ,

where Q = 2 n + 2 is the so-called homogeneous dimension of ℍ n .

The Heisenberg group is a homogeneous group with the norm:

| x | h = [ ( ∑ i = 1 2 n x i 2 ) 2 + x 2 n + 1 2 ] 1 / 4 ,

and the Heisenberg distance d, generated by this norm is given by

d ( p , q ) = d ( q − 1 p , 0 ) = | q − 1 p | h .

Note that d satisfies triangular inequality and is left-invariant in the sense that

d ( r ⋅ p , r ⋅ q ) = d ( p , q ) , ∀ p , q , r ∈ ℍ n .

The ball and sphere on ℍ n , for r > 0 and x ∈ ℍ n , can be defined as

B ( x , r ) = { y ∈ ℍ n : d ( x , y ) < r }

and

S ( x , r ) = { y ∈ ℍ n : d ( x , y ) = r } ,

respectively. To compute the measure of this ball on ℍ n , we proceed as follows:

| B ( x , r ) | = | B ( 0 , r ) | = Ω Q r Q ,

where Ω Q , being a function of n only, is the volume of the unit ball B ( 0 , 1 ) . Also, the area of unit sphere S ( 0 , 1 ) on ℍ n is w Q = Q Ω Q . For further readings on the Heisenberg group, we refer the interested reader to the book by Folland and Stein [28] and previous studies [29,30,31].

Now, we are in position to define the Hausdorff operator and its commutators on the Heisenberg group ℍ n . Let Φ be a locally integrable function on ℍ n . The Hausdorff operators on ℍ n are defined by:

T Φ f ( x ) = ∫ ℍ n Φ ( y ) | y | h Q f ( δ | y | h − 1 x ) d y , T Φ , A f ( x ) = ∫ ℍ n Φ ( y ) | y | h Q f ( A ( y ) x ) d y ,

where A ( y ) is a matrix-valued function, and we assume that det A ( y ) ≠ 0 almost everywhere in the support of Φ . Also, we define the commutators T Φ , A b of T Φ , A with locally integrable function b as

(1.5) T Φ , A b ( f ) = b T Φ , A ( f ) − T Φ , A ( b f ) .

In this article, we will study the boundedness of T Φ , A b on the weighted Herz spaces K ̇ q 2 α 2 , p ( ℍ n ; w ) , defined in Section 2, with the Heisenberg group as underlying space. Section 2 contains some basic definitions and notations likewise some necessary propositions which will be used in the succeeding sections. Finally, Section 3 is reserved for the main results of this study along with their proofs.

2 Some definitions and notations

In 1972, Muckenhoupt [32] studied the Hardy-Littlewood maximal function on weighted L ^p spaces and introduced the theory of A _p weights as a result. The theory was well studied in the later work by García-Cuerva and Rubio de Francia [33]. An extension of this theory, in the settings of the Heisenberg group ℍ n , was provided in [29] and studied in [30,31]. Any non-negative, locally integrable function w on ℍ n can be given the role of a weight. The notation w(E) serves to define weighted measure of E ⊂ ℍ n , that is, w ( E ) = ∫ E w ( x ) d x . Also, if p and p′ satisfy 1 / p + 1 / p ′ = 1 , then they will be called mutually conjugate indices. Next, let us recall some basic definitions and properties of A _p weights on the Heisenberg group which will be used in the sequel.

Definition 2.1

We say that w belongs to the Muckenhoupt class A p ( ℍ n ) , 1 < p < ∞ , if there exists a C > 0 such that for every ball B ⊂ ℍ n ,

( 1 | B | ∫ B w ( x ) d x ) ( 1 | B | ∫ B w ( x ) − p ′ / p d x ) p / p ′ ≤ C .

Also, w ∈ A 1 if there exists a constant C > 0 such that for every ball B ⊂ ℍ n ,

( 1 | B | ∫ B w ( x ) d x ) ≤ C essinf x ∈ B w ( x ) .

When p = ∞ , we define A ∞ = ∪ 1 ≤ p < ∞ A p .

According to Proposition 2.2 in [25], we have A p ( ℍ n ) ⊂ A q ( ℍ n ) , for 1 ≤ p < q < ∞ , and if w ∈ A p ( ℍ n ) , 1 < p < ∞ , then there is an ε > 0 such that p − ε > 1 and w ∈ A p − ε ( ℍ n ) . Therefore, we may use q w ≔ inf { q > 1 : w ∈ A q } to denote the critical index of w.

Definition 2.2

We say that w belongs to the reverse Hölder class R H r ( ℍ n ) , if there exists a fixed constant C > 0 and r > 1 , such that for every ball B ⊂ ℍ n ,

( 1 | B | ∫ B w r ( x ) d x ) 1 / r ≤ C | B | ∫ B w ( x ) d x .

In [31], it was proved that w ∈ A ∞ ( ℍ n ) if and only if there exist some r > 1 such that w ∈ R H r ( ℍ n ) . In addition, if w ∈ R H r ( ℍ n ) , r > 1 , then for some ε > 0 we have w ∈ R H r + ε ( ℍ n ) . We therefore use r w ≔ sup { r > 1 : w ∈ R H r ( ℍ n ) } to denote the critical index of w for the reverse Hölder condition.

A particular case of Muckenhoupt A p ( ℍ n ) weights is the power weight function | x | h α . From Proposition 2.3 in [25], for x ∈ ℍ n , we have | x | h α ∈ A 1 ( ℍ n ) if and only if − Q < α ≤ 0 . Also, for 1 < p < ∞ , | x | h ∈ A p ( ℍ n ) , if and only if − Q < α < Q ( p − 1 ) . In view of these observations, it is easy to see that for 0 < α < ∞ ,

| x | h α ∈ ⋂ Q + α Q < p < ∞ A p ( ℍ n ) ,

where ( Q + α ) / Q is known as the critical index of | x | h α .

The following two Propositions, proved in [25], concerning A p ( ℍ n ) weights will be useful in establishing weighted estimates for T Φ, A b on Herz-type spaces on ℍ n .

Proposition 2.3

Let w ∈ A p ∩ R H r ( ℍ n ) , p ≥ 1 and r > 1 . Then, there exist constants C 1 , C 2 > 0 such that

C 1 ( | E | | B | ) p ≤ w ( E ) w ( B ) ≤ C 2 ( | E | | B | ) ( r − 1 ) / r ,

for any measurable subset E of a ball B. In general, for any λ > 1 ,

w ( B ( x 0 , λ R ) ≤ λ Q p w ( B ( x 0 , R ) ) .

Proposition 2.4

If w ∈ A p ( ℍ n ) , 1 ≤ p < ∞ , then for any f ∈ L loc 1 ( ℍ n ) and any ball B ⊂ ℍ n

1 | B | ∫ B | f ( x ) | d x ≤ C ( 1 w ( B ) ∫ B | f ( x ) | p w ( x ) d x ) 1 / p .

For any measurable set E ⊂ ℍ n , the weighted Lebesgue space L p ( E ; w ) is the space of all functions f satisfying the norm condition

∥ f ∥ L p ( E ; w ) = ( ∫ E | f ( x ) | p w ( x ) d x ) 1 / p < ∞ ,

where 1 ≤ p < ∞ and w is a weight function on ℍ n . When p = ∞ , we have L ∞ ( ℍ n ; w ) = L ∞ ( ℍ n ) and ∥ f ∥ L ∞ ( ℍ n ; w ) = ∥ f ∥ L ∞ ( ℍ n ) .

Let B k ≔ { x ∈ ℍ n : | x | h < 2 k } , E k = B k / B k − 1 for k ∈ Z . Then, the homogeneous weighted Herz space in the setting of the Heisenberg group can be defined as follows.

Definition 2.5

[25] Let α ∈ ℝ , 0 < p , q < ∞ , and w is a weight function on ℍ n . The homogeneous weighted Herz space K ̇ q α , p ( ℍ n ) is defined by

K ̇ q α , p ( ℍ n ; w ) ≔ { f ∈ L loc q ( ℍ n / { 0 } ; w ) : ∥ f ∥ K ̇ q α , p ( ℍ n ; w ) < ∞ } ,

where

∥ f ∥ K ̇ q α , p ( ℍ n ; w ) = { ∑ k = − ∞ ∞ w ( B k ) α p / Q ∥ f ∥ L p ( E k ; w ) p } 1 / p .

When w = 1 , we obtain K ̇ q α , p ( ℍ n ) introduced in [34]. It is easy to verify that K ̇ p α , p ( ℍ n ) = L p ( ℍ n , | ⋅ | h α p ) . Hence, Herz space can be considered as an extension of power weighted Lebesgue space. Some relevant papers on Herz-type spaces and Hardy spaces associated with them along with their application include [35,36,37,38,39,40,41,42,43].

Definition 2.6

[44] Let 1 < q < ∞ and w be a weight function on ℍ n . Then, we say a function f ∈ L loc q ( ℍ n ; w ) belongs to the weighted central bounded mean oscillation (CBMO) space C M ̇ O q ( ℍ n ; w ) if

∥ f ∥ C M ̇ O q ( ℍ n ; w ) = sup R > 0 ( 1 w ( B ( 0 , R ) ) ∫ B ( 0 , R ) | f ( x ) − f B | q w ( x ) ) 1 / q < ∞ ,

where

(2.1) f B = 1 | B ( 0 , r ) | ∫ B ( 0 , r ) f ( x ) d x .

For detailed study of CBMO space on ℝ n , we refer the reader to [45,46].

Recently, weighted boundedness of matrix Hausdorff operators and their commutators defined on different underlying spaces are reported in [44,47,48,49,50,51,52,53].

Lemma 2.7

[25] Suppose that the ( 2 n + 1 ) × ( 2 n + 1 ) matrix M is invertible. Then,

(2.2) ∥ M ∥ − Q ≤ | det M − 1 | ≤ ∥ M − 1 ∥ Q ,

where

(2.3) ∥ M ∥ = sup x ∈ ℍ n , x ≠ 0 | M x | h | x | h .

Also, when A _p weights are reduced to the power function, we shall use the notation v ( ⋅ ) instead of w ( ⋅ ) , that is, v ( ⋅ ) = | ⋅ | h β . In that case, an easy computation results in:

(2.4) v ( B k ) = ∫ | x | h ≤ 2 k | x | h β d x = ω Q 2 k ( Q + β ) / ( β + Q ) .

Moreover, in the case of boundedness of T Φ , A b on the power-weighted Herz space, we shall frequently use the piecewise defined function G:

G ( M , δ β ) = { ∥ M ∥ δ β if β > 0 , ∥ M − 1 ∥ − δ β if β ≤ 0 ,

where M is any invertible matrix, α ∈ ℝ and δ is a positive real number. Then, it is easy to see that G satisfies:

(2.5) G ( M , β ( 1 / q + 1 / p ) ) = G ( M , β / q ) G ( M , β / p ) ,

where p , q ∈ Z + .

Proposition 2.8

Suppose that the ( 2 n + 1 ) × ( 2 n + 1 ) matrix M is invertible. Let β > − n , v ( x ) = | x | h β and x ∈ ℍ n , then

v ( M x ) ≤ { ∥ M ∥ β v ( x ) if β > 0 , ∥ M − 1 ∥ − β v ( x ) if β ≤ 0 , = G ( M , β ) v ( x ) .

From this point forward, the notations A ≼ B will imply that A ≤ C B , for some C > 0 . Similarly, for some positive constants C 1 and C 2 , if A ≤ C 1 B and B ≤ C 2 A , then we will write A ≃ B . Also, we set λ B ( 0 , R ) = B ( 0 , λ R ) , for λ > 0 .

3 Main results and their proofs

This section contains the main results of this study and the relevant proofs. Our first result is as follows.

Theorem 3.1

Let 1 ≤ p , q , q 1 , q 2 ≤ ∞ and α 1 , α 2 ∈ ℝ with α 1 < 0 . Suppose that 1 / s = 1 / q 1 + 1 / q and α 1 / Q + 1 / q 1 = α 2 / Q + 1 / q 2 . In addition, let w ∈ A 1 with the critical index r w for the reverse Hölder condition and s > q 2 r w / ( r w − 1 ) .

If 1 / q 1 + α 1 / Q ≥ 0 , then for any 1 < δ < r w ,
∥ T Φ , A b f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) ≤ K 1 ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; w ) ,
where
K 1 = ∫ ∥ A ( y ) ∥ < 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ − α 1 log 2 ∥ A ( y ) ∥ d y + ∫ ∥ A ( y ) ∥ ≥ 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ Q / q 1 − ( α 1 + Q / q 1 ) ( δ − 1 ) / δ log 2 ∥ A ( y ) ∥ d y ;
If α 1 / Q + 1 / q 1 < 0 , then for any 1 < δ < r w

∥ T Φ , A b ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) ≤ K 2 ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; w ) ,

where

K 2 = ∫ ∥ A ( y ) ∥ ≥ 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ − α 1 log 2 ∥ A ( y ) ∥ d y + ∫ ∥ A ( y ) ∥ < 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ Q / q 1 − ( α 1 + Q / q 1 ) ( δ − 1 ) / δ log 2 ∥ A ( y ) ∥ d y .

When general weights are reduced to power weights, then the next theorem is as follows.

Theorem 3.2

Let 1 ≤ p < ∞ , 1 < q , q 1 , q 2 < ∞ and β > − n . If 1 / q 2 = 1 / q + 1 / q 1 and 1 / q + α 2 / Q = α 1 / Q , then we have

∥ T Φ , A b ∥ K ̇ q 2 α 2 , p ( ℍ n ; v ) ≼ K 3 ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; v ) ,

where K 3 is

K 3 = { ∫ ℍ n Θ ( y ) ( 1 + log 2 ( ∥ A − 1 ( y ) ∥ ∥ A ( y ) ∥ ) ) d y , if α 1 = 0 , ∫ ℍ n Θ ( y ) G ( A − 1 ( y ) , α 1 ( Q + β ) / Q ) d y , if α 1 ≠ 0 ,

and

Θ ( y ) = | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 ( log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ < 1 } + log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ ≥ 1 } ) × G ( A − 1 ( y ) , β / q 1 ) ( 1 + | det A − 1 ( y ) | 1 / q G ( A − 1 ( y ) , β / q ) ∥ A ( y ) ∥ ( Q + β ) / q ) .

3.1 Proof of Theorem 3.1

Here, we have to show that

{ ∑ k = − ∞ ∞ w ( B k ) α 2 p / Q ∥ T Φ , A b f ∥ L q 2 ( E k , w ) p } 1 / p ≼ ∥ f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) .

By the Minkowski inequality and necessary splitting, an upper bound for the inner norm ∥ T Φ , A b f ∥ L q 2 ( E k , w ) p can be obtained as:

(3.1) ∥ ( T Φ , A b f ) ∥ L q 2 ( E k ; w ) = ∥ ( ∫ ℍ n Φ ( y ) | y | h Q ( b ( x ) − b ( A ( y ) x ) ) f ( A ( y ) x ) d y ) ∥ L q 2 ( E k ; w ) ≤ ∫ ℍ n Φ ( y ) | y | h Q ∥ ( b ( x ) − b ( A ( y ) x ) ) f ( A ( y ) x ) ∥ L q 2 ( E k ; w ) d y ≤ ∫ ℍ n Φ ( y ) | y | h Q ∥ ( b ( x ) − b B k ) f ( A ( y ) x ) ∥ L q 2 ( E k ; w ) d y + ∫ ℍ n Φ ( y ) | y | h Q ∥ ( b ( A ( y ) x ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) x ) ∥ L q 2 ( E k ; w ) d y + ∫ ℍ n Φ ( y ) | y | h Q ∥ ( b B k − b ∥ A ( y ) ∥ B k ) f ( A ( y ) x ) ∥ L q 2 ( E k ; w ) d y = I 1 + I 2 + I 3 .

While targeting I ₁, we first compute ∥ ( b ( x ) − b ( A ( y ) x ) ) f ( A ( y ) x ) ∥ L q 2 ( E k ; w ) . The condition s > q 2 r w / ( r w − 1 ) implies that there exist 1 < r < r w such that s = q 2 r ′ . Therefore, by the Hölder inequality and the reverse Hölder condition, we have

(3.2) ∥ ( b ( ⋅ ) − b B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; w ) = ( ∫ E k | ( b ( x ) − b B k ) f ( A ( y ) x ) | s d x ) 1 / s ( ∫ E k w ( x ) r d x ) 1 / r q 2 ≼ | B k | − 1 / s w ( B k ) 1 / q 2 ∥ ( b ( ⋅ ) − b B k ) f ( A ( y ) ⋅ ) ∥ L s ( E k ) .

Next, using the condition 1 / s = 1 / q + 1 / q 1 , we can have

(3.3) ∥ ( b ( ⋅ ) − b B k ) f ( A ( y ) ⋅ ) ∥ L s ( E k ) ≤ ∥ b ( ⋅ ) − b B k ∥ L q ( B k ) ∥ f ( A ( y ) ⋅ ) ∥ L q 1 ( B k ) .

In second factor, on the right side of inequality (3.3), a change of variables along with Proposition 2.4 yields

Similarly, the other factor on the right hand of inequality (3.3), in view of Proposition 2.4, gives

(3.5) ∥ b ( ⋅ ) − b B k ∥ L q ( B k ) ≼ | B k | 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; w ) .

Inequalities (3.2)–(3.5) together yield

∥ ( b ( ⋅ ) − b B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; w ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 w ( B k ) 1 / q 2 w ( ∥ A ( y ) ∥ B k ) 1 / q 1 .

Hence, we obtain the following estimate for I 1 :

I 1 ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ℍ n | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 w ( B k ) 1 / q 2 w ( ∥ A ( y ) ∥ B k ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) d y .

Next, we fix to estimate I 2 , which is given by

I 2 = ∫ ℍ n | Φ ( y ) | | y | h Q ∥ ( b ( A ( y ) ⋅ ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; w ) d y .

Since s = q 2 r ′ , therefore, we infer from (3.2) that

(3.6) ∥ ( b ( A ( y ) ⋅ ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; w ) ≼ | B k | − 1 / s w ( B k ) 1 / q 2 ∥ ( b ( A ( y ) ⋅ ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) ⋅ ) ∥ L s ( E k ) .

Applying the change of variables formula, Proposition 2.4 and Hölder’s inequality, we have

By virtue of (3.6) and (3.7), the expression for I ₂ assumes the following form:

I 2 ≼ ∥ b ∥ C M ̇ O q ( ℍ n , w ) ∫ ℍ n | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / s w ( B k ) 1 / q 2 w ( ∥ A ( y ) ∥ B k ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) d y .

Now, the estimation of I 3 , given by

I 3 = ∫ ℍ n | Φ ( y ) | | y | h Q ∥ f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ) | b B k − b ∥ A ( y ) ∥ B k | d y ,

requires the bounds for ∥ f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ) and | b B k − b ∥ A ( y ) ∥ B k | . First, we consider ∥ f ( A ( y ) ⋅ ) ∥ L q 2 ( E k , w ) . In view of the condition s = q 2 r ′ , we use the Hölder inequality and the reverse Hölder condition to obtain

Furthermore, the condition 1 / s = 1 / q + 1 / q 1 and inequality (3.4) help us to write

(3.9) ∥ f ( A ( y ) ⋅ ) ∥ L s ( B k ) = | B k | 1 / q ∥ f ( A ( y ) ⋅ ) ∥ L q 1 ( B k ) ≼ | B k | 1 / s ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 w ( ∥ A ( y ) ∥ B k ) − 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) .

We combine inequalities (3.8) and (3.9) to substitute the result in the expression for I 3 , which now becomes

I 3 ≼ ∫ ℍ n | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 w ( B ( 0 , 2 k ) ) 1 / q 2 w ( ∥ A ( y ) ∥ ( B k ) ) 1 / q 1 ∥ f ∥ L q 1 ( w ( ∥ A ( y ) ∥ B k ) , w ) | b B k − b ∥ A ( y ) ∥ B k | d y .

Now, it turns to bound | b B k − b ∥ A ( y ) ∥ B k | . For this purpose, we split the integral as follows:

I 3 ≼ ∫ ∥ A ( y ) ∥ < 1 | b B k − b ∥ A ( y ) ∥ B k | Ψ ( y ) d y + ∫ ∥ A ( y ) ∥ ≥ 1 | b B k − b ∥ A ( y ) ∥ B k | Ψ ( y ) d y = I 31 + I 32 ,

where, for the convenience’s sake, we used the following notation:

Ψ ( y ) = | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 w ( B ( 0 , 2 k ) ) 1 / q 2 w ( ∥ A ( y ) ∥ ( B k ) ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) .

Further decomposition of integral for I ₃₁ results in:

I 31 = ∑ j = 0 ∞ ∫ 2 − j − 1 ≤ ∥ A ( y ) ∥ < 2 − j Ψ ( y ) { ∑ i = 1 j | b 2 − 1 B k − b 2 − i + 1 B k | + | b 2 − j B k − b ∥ A ( y ) ∥ B k | } d y .

The first term inside the curly brackets can be approximated using Proposition 2.4, that is,

Similarly, for the second term inside the curly brackets in the expression of I 31 , we have

| b 2 − j B k − b ∥ A ( y ) ∥ B k | ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) .

Therefore, we finish the estimation of I ₃₁ by writing

I 31 ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∑ j = 0 ∞ ∫ 2 − j − 1 ≤ ∥ A ( y ) ∥ < 2 − j Ψ ( y ) ( j + 1 ) d y ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ∥ A ( y ) ∥ < 1 Ψ ( y ) log 2 ∥ A ( y ) ∥ d y .

In a similar fashion, the integral I ₃₂ gives us

I 32 = ∫ ∥ A ( y ) ∥ ≥ 1 Ψ ( y ) | b B k − b ∥ A ( y ) ∥ B k | d y = ∑ j = 0 ∞ ∫ 2 j ≤ ∥ A ( y ) ∥ < 2 j + 1 Ψ ( y ) { ∑ i = 1 j | b 2 i B k − b 2 i + 1 B k | + | b 2 j + 1 B k − b ∥ A ( y ) ∥ B k | } d y ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ∥ A ( y ) ∥ ≥ 1 Ψ ( y ) log 2 ∥ A ( y ) ∥ d y .

A combination of expressions for I ₁, I ₂, I ₃₁ and I ₃₂ gives

∥ T Φ , A b f ∥ L q 2 ( E k ; w ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ℍ n | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 × ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) w ( B ( 0 , 2 k ) ) 1 / q 2 w ( ∥ A ( y ) ∥ B k ) ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) × max { log 2 ∥ A ( y ) ∥ , log ( 2 ∥ A ( y ) ∥ ) } d y .

Keeping in view the definition of the Herz space, factors containing the index k in the expression of Ψ ( y ) are important. Therefore, to proceed further and to avoid repetition of unimportant factors relative to the Herz space, we have to modify and rename the expression for Ψ . Hence, in the remaining of this paper we shall use the following notation:

Ψ ˜ ( y ) = | Φ ( y ) | | y | h Q ( | det A − 1 ( y ) | ∥ A ( y ) ∥ Q ) 1 / q 1 ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) max { log 2 ∥ A ( y ) ∥ , log 2 ∥ A ( y ) ∥ } .

Then,

∥ T Φ , A b f ∥ L q 2 ( E k ; w ) ≤ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ℍ n Ψ ˜ ( y ) w ( B k ) 1 / q 2 w ( ∥ A ( y ) ∥ B k ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) d y .

Finally, we take into consideration the definition of Herz space and employ the Minkowski inequality to have

(3.10) ∥ T Φ , A b f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) = { ∑ k = − ∞ ∞ w ( B k ) α 2 p Q ∥ T Φ , A b f ∥ L q 2 ( E k ; w ) p } 1 / p ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∫ ℍ n Ψ ˜ ( y ) { ( ∑ k = − ∞ ∞ w ( B k ) α 2 / Q + 1 / q 2 w ( ∥ A ( y ) ∥ B k ) 1 / q 1 ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ B k ; w ) ) p } 1 / p d y .

Comparing inequality (3.10) with inequality (3.9) in [25], we found that the term inside the curly brackets is same in both these inequalities, the only difference lies in the integrands outside the curly brackets along with a constant multiple ∥ b ∥ C M ̇ O q ( ℍ n ; w ) outside the integral. Therefore, inequality (3.10) can be written as:

(3.11) ∥ T Φ , A b f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∑ j = − ∞ ∞ ∫ 2 j − 1 < ∥ A ( y ) ∥ ≤ 2 j Ψ ˜ ( y ) { ∑ k = − ∞ ∞ [ ( w ( B k ) w ( B k + j ) ) α 1 / Q + 1 / q 1 × ∑ l = − ∞ j ( w ( B k + j ) w ( B k + l ) ) α 1 / Q w ( B k + l ) α 1 / Q ∥ f ∥ L q 1 ( E k + l ; w ) ] p } 1 / p d y ,

where the condition α 1 / Q + 1 / q 1 = α 2 / Q + 1 / q 2 is utilized in obtaining the last inequality.

Under the stated condition that α 1 < 0 and l ≤ j , we use Proposition 2.3 to have

(3.12) ( w ( B k + j ) w ( B k + l ) ) α 1 / Q ≼ ( | B k + j | | B k + l | ) α 1 ( δ − 1 ) / ( Q δ ) = 2 ( j − l ) α 1 ( δ − 1 ) / δ ,

for any 1 < δ < r w .

In view of Proposition 2.3, if α 1 / Q + 1 / q 1 ≥ 0 , then

(3.13) ( w ( B k ) w ( B k + j ) ) α 1 / Q + 1 / q 1 ≼ { 2 − j Q ( α 1 / Q + 1 / q 1 ) , if j ≤ 0 , 2 − j Q ( α 1 / Q + 1 / q 1 ) ( δ − 1 ) / δ , if j > 0 ,

and if α 1 / Q + 1 / q 1 < 0 , then

(3.14) ( w ( B k ) w ( B k + j ) ) α 1 / Q + 1/ q 1 ≼ { 2 j Q ( α 1 / Q + 1/ q 1 ) ( δ − 1)/ δ , if j ≤ 0, 2 j Q ( α 1 / Q + 1/ q 1 ) , if j > 0,

for any 1 < δ < r w .

Thus, for α 1 / Q + 1 / q 1 ≥ 0 , from inequalities (3.11)–(3.13), for any 1 < δ < r w , we have

∥ T Φ , A b f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∑ j = − ∞ 0 ∫ 2 j − 1 < ∥ A ( y ) ∥ ≤ 2 j Ψ ˜ ( y ) ∥ A ( y ) ∥ − α 1 − Q / q 1 × ∑ l = − ∞ j 2 α 1 ( j − l ) ( δ − 1 ) / δ { ∑ k = − ∞ ∞ w ( B k + l ) α 1 p / Q ∥ f ∥ L q 1 ( E k + l ; w ) p } 1 / p d y + ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∑ j = 1 ∞ ∫ 2 j − 1 < ∥ A ( y ) ∥ ≤ 2 j Ψ ˜ ( y ) ∥ A ( y ) ∥ ( α 1 + Q / q 1 ) ( δ − 1 ) / δ × ∑ l = − ∞ j 2 α 1 ( j − l ) ( δ − 1 ) / δ { ∑ k = − ∞ ∞ w ( B k + l ) α 1 p / Q ∥ f ∥ L q 1 ( E k + l ; w ) p } 1 / p d y .

Replacing Ψ ˜ ( y ) with its value in the above inequality, we get

∥ T Φ , A b f ∥ K ̇ q 2 α 2 , p ( ℍ n ; w ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; w ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; w ) × { ∫ ∥ A ( y ) ∥ < 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ − α 1 log 2 ∥ A ( y ) ∥ d y + ∫ ∥ A ( y ) ∥ ≥ 1 | Φ ( y ) | | y | h Q ( 1 + | det A − 1 ( y ) | 1 / q ∥ A ( y ) ∥ Q / q ) × | det A − 1 ( y ) | 1 / q 1 ∥ A ( y ) ∥ Q / q 1 − ( α 1 + Q / q 1 ) ( δ − 1 ) / δ log ( 2 ∥ A ( y ) ∥ ) d y } .

This completes the proof of Theorem 3.1(i).

Similarly, when α 1 / Q + 1 / q 1 < 0 , by using inequalities (3.11), (3.12) and (3.14), the second part of Theorem 3.1 can be proved easily. Hence, we complete the proof of Theorem 3.1.

3.2 Proof of Theorem 3.2

Following the proof of Theorem 3.1, we write:

∥ T Φ , A b ∥ L q 2 ( E k ; v ) ≤ J 1 + J 2 + J 3 ,

where J 1 , J 2 and J 3 are similar to I 1 , I 2 and I 3 in the previous theorem with w ( ⋅ ) replaced by v ( ⋅ ) = | ⋅ | h α . Then, by using the Hölder inequality and change of variables, we obtain

J 1 ≤ ∫ ℍ n | Φ ( y ) | | y | h Q ( ∫ E k | b ( x ) − b B k | q v ( x ) d x ) 1 / q ( ∫ E k | f ( A ( y ) x ) | q 1 v ( x ) d x ) 1 / q 1 d y ≤ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 ( ∫ A ( y ) E k | f ( z ) | q 1 v ( A − 1 ( y ) z ) d z ) 1 / q 1 d y .

Using Proposition 2.8, we get

J 1 ≤ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 ( ∫ A ( y ) E k | f ( x ) | q 1 G ( A − 1 ( y ) , β / q 1 ) v ( x ) d x ) 1 / q 1 d y ≤ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 G ( A − 1 ( y ) , β / q 1 ) ∥ f ∥ L q 1 ( ∥ A ( y ) ∥ E k ; v ) d y .

Next, the expression for J 2 can be written as:

(3.15) J 2 = ∫ ℍ n | Φ ( y ) | | y | h Q ∥ ( b ( A ( y ) ⋅ ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; v ) d y .

Changing variables and using the condition q 2 / q + q 2 / q 1 = 1 , we get

(3.16) ∥ ( b ( A ( y ) x ) − b ∥ A ( y ) ∥ B k ) ( f ( A ( y ) . ) ) ∥ L q 2 ( E k ; v ) = ( ∫ E k | ( b ( A ( y ) x ) − b ∥ A ( y ) ∥ B k ) f ( A ( y ) x ) | q 2 v ( x ) d x ) 1 / q 2 = | det A − 1 ( y ) | 1 / q 2 G ( A − 1 ( y ) , β / q 2 ) ( ∫ A ( y ) E k | ( b ( x ) − b ∥ A ( y ) ∥ B k ) f ( x ) | q 2 v ( x ) d x ) 1 / q 2 ≤ | det A − 1 ( y ) | 1 / q 2 G ( A − 1 ( y ) , β / q 2 ) ( ∫ ∥ A ( y ) ∥ B k | b ( x ) − b ∥ A ( y ) ∥ B k | q v ( x ) d x ) 1 / q ( ∫ A ( y ) E k | f ( x ) | q 1 v ( x ) d x ) 1 / q 1 = | det A − 1 ( y ) | 1 / q 2 G ( A − 1 ( y ) , β / q 2 ) v ( ∥ A ( y ) ∥ B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∥ f ∥ L q 1 ( A ( y ) E k ; v ) .

It is easy to see that v ( ∥ A ( y ) ∥ B k ) = ∥ A ( y ) ∥ Q + β v ( B k ) . Using properties (2.5) and (3.16), inequality (3.15) becomes:

J 2 = v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 2 G ( A − 1 ( y ) , β / q ) G ( A − 1 ( y ) , β / q 1 ) ∥ A ( y ) ∥ ( Q + β ) / q ∥ f ∥ L q 1 ( A ( y ) E k ; v ) d y .

It remains to estimate J 3 . A change of variables following the Hölder inequality and Proposition 2.8 gives us

J 3 = ∫ ℍ n | Φ ( y ) | | y | h Q ∥ ( b B k − b ∥ A ( y ) ∥ B k ) f ( A ( y ) ⋅ ) ∥ L q 2 ( E k ; v ) d y = ∫ ℍ n | Φ ( y ) | | y | h Q ∥ f ( A ( y ) x ) ∥ L q 2 ( E k ; v ) | b B k − b ∥ A ( y ) ∥ B k | d y ≤ ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 2 G ( A − 1 ( y ) , β / q 2 ) v ( ∥ A ( y ) ∥ B k ) 1 / q ∥ f ∥ L q 1 ( A ( y ) E k ; v ) | b B k − b ∥ A ( y ) ∥ B k | d y .

Next, if ∥ A ( y ) ∥ < 1 , then there exists an integer j ≥ 0 , such that

2 − j − 1 ≤ ∥ A ( y ) ∥ < 2 − j .

Therefore,

Similarly, for ∥ A ( y ) ∥ ≥ 1 , we have

| b B k − b ∥ A ( y ) ∥ B k | ≤ ∥ b ∥ C M ̇ O q ( ℍ n ; v ) log 2 ∥ A ( y ) ∥ .

Hence,

J 3 ≼ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 G ( A − 1 ( y ) , β / q 2 ) × G ( A − 1 ( y ) , β / q ) ∥ A ( y ) ∥ ( Q + β ) / q ( log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ < 1 } + log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ ≥ 1 } ) ∥ f ∥ L q 1 ( A ( y ) E k ; v ) d y .

Thus, combining J ₁, J ₂ and J ₃, we get

(3.17) ∥ T Φ , A b ∥ L q 2 ( E k ; v ) ≼ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n | Φ ( y ) | | y | h Q | det A − 1 ( y ) | 1 / q 1 × G ( A − 1 ( y ) , β / q 1 ) ( 1 + | det A − 1 ( y ) | 1 / q G ( A − 1 ( y ) , β / q ) ∥ A ( y ) ∥ ( Q + β ) / q ) × ( log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ < 1 } + log 2 ∥ A ( y ) ∥ χ { ∥ A ( y ) ∥ ≥ 1 } ) ∥ f ∥ L q 1 ( A ( y ) E k ; v ) d y .

For the approximation of ∥ f ( ⋅ ) ∥ L q ( A ( y ) C k ) , we consider the method used in [25]. Hence, the definition of E k and (2.2) imply that

A ( y ) E k ⊂ { x : ∥ A − 1 ( y ) ∥ − 1 2 k − 1 < | x | h < ∥ A ( y ) ∥ 2 k } .

Now, there exists an integer l such that for any y ∈ supp ( Φ ) , we have

(3.18) 2 l < ∥ A − 1 ( y ) ∥ − 1 < 2 l + 1 .

Finally, the inequality ∥ A − 1 ( y ) ∥ − 1 ≤ ∥ A ( y ) ∥ implies that there exists a non-negative integer m satisfying:

(3.19) 2 l + m < ∥ A ( y ) ∥ < 2 l + m + 1 .

We infer from (3.18) and (3.19) that:

log 2 ( ∥ A ( y ) ∥ ∥ A − 1 ( y ) ∥ / 2 ) < m < log 2 ( 2 ∥ A ( y ) ∥ ∥ A − 1 ( y ) ∥ ) .

Therefore,

A ( y ) E k ⊂ { x : 2 l + k − 1 < | x | h < 2 k + l + m + 1 } .

Hence,

(3.20) ∥ f ∥ L q 1 ( A ( y ) E k ; v ) ≤ ∑ j = l l + m + 1 ∥ f ∥ L q 1 ( E k + j ; v ) .

Incorporating inequality (3.20) into inequality (3.17), we obtain

(3.21) ∥ T Φ , A b ∥ L q 2 ( E k ; v ) ≤ v ( B k ) 1 / q ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n Θ ( y ) ∑ j = l l + m + 1 ∥ f ∥ L q 1 ( E k + j ; v ) d y ,

where

Using the Minkowski inequality and the condition 1 / q + α 2 / Q = α 1 / Q yields

∥ T Φ , A b ∥ K ̇ q 2 α 2 , p ( ℍ n ; v ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; v ) { ∑ k = − ∞ ∞ ( v ( B k ) 1 / q + α 2 / Q ∫ ℍ n Θ ( y ) ∑ j = l l + m + 1 ∥ f ∥ L q 1 ( E k + j ; v ) d y ) p } 1 / p ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∫ ℍ n Θ ( y ) ∑ j = l l + m + 1 v ( B − j ) α 1 / Q { ∑ k = − ∞ ∞ ( v ( B k + j ) α 1 / Q ∥ f ∥ L q 1 ( E k + j ; v ) ) p } 1 / p d y ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; v ) ∫ ℍ n Θ ( y ) ∑ j = l l + m + 1 v ( B − j ) α 1 / Q d y .

It is easy to see that

∑ j = l l + m + 1 v ( B − j ) α 1 / Q ≃ ∑ j = l l + m + 1 2 − j α 1 ( Q + β ) / Q .

Next, for α 1 = 0 ,

∑ j = l l + m + 1 2 − j α 1 ( Q + β ) / Q = m + 2 ≼ 1 + log 2 ( ∥ A − 1 ( y ) ∥ ∥ A ( y ) ∥ ) ,

and for α 1 ≠ 0 ,

∑ j = l l + m + 1 2 − j α 1 ( Q + β ) / Q ≃ 2 − l α 1 ( Q + β ) / Q ≼ { ∥ A − 1 ( y ) ∥ α 1 ( Q + β ) / Q , if α 1 > 0 , ∥ A ( y ) ∥ − α 1 ( Q + β ) / Q , if α 1 < 0 , = G ( A − 1 ( y ) , α 1 ( Q + β ) / Q ) .

Therefore,

∥ T Φ , A b ∥ K ̇ q 2 α 2 , p ( ℍ n ; v ) ≼ ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; v ) { ∫ ℍ n Θ ( y ) ( 1 + log 2 ( ∥ A − 1 ( y ) ∥ ∥ A ( y ) ∥ ) ) d y , if α 1 = 0 , ∫ ℍ n Θ ( y ) G ( A − 1 ( y ) , α 1 ( Q + β ) / Q ) d y , if α 1 ≠ 0 , = K 3 ∥ b ∥ C M ̇ O q ( ℍ n ; v ) ∥ f ∥ K ̇ q 1 α 1 , p ( ℍ n ; v ) .

Thus, we complete the proof of Theorem 3.2.

4 Conclusion

In this article, the authors obtained the boundedness of the commutator of the matrix Hausdorff operator on the homogeneous weighted Herz space in the settings of the Heisenberg group. As an application, the authors also investigated the particular case of Muckenhoupt A _p weights, namely, the power weights. Some potential directions for the future works include the boundedness of the same operator on homogeneous weighted Herz-Morrey spaces and weighted Herz-type Hardy spaces defined on homogeneous groups.

Acknowledgments

The authors would like to thank the referees for careful reading of the paper and valuable suggestions. Amjad Hussain is supported by Higher Education Commission (HEC) of Pakistan through the National Research Program for Universities (NRPU) Project No: 7098/Federal/NRPU/R&D/HEC/2017 and the Quaid-I-Azam University Research Fund (URF) Project.

Conflicts of interest: The authors declare that they have no competing interest.

References

[1] A. K. Lerner and E. Liflyand, Multidimensional Hausdorff operator on the real Hardy space, J. Austr. Math. Soc. 83 (2007), 65–72.10.1017/S1446788700036399Search in Google Scholar

[2] E. Liflyand, Boundeness of multidimensional Hausdorff operators on H1(ℝn), Acta Sci. Math. 74 (2008), 845–851.Search in Google Scholar

[3] E. Liflyand and A. Miyachi, Boundedness of multidimensional Hausdorff operators in Hp spaces, 0 < p < 1, Trans. Amer. Math. Soc. 371 (2018), 4793–4814.10.1090/tran/7572Search in Google Scholar

[4] J. C. Chen, D. S. Fan, and J. Li, Hausdorff operators on function spaces, Chin. Ann. Math. 33 (2012), 537–556.10.1007/s11401-012-0724-1Search in Google Scholar

[5] J. C. Chen, D. S. Fan, and C. J. Zand, Boundedness of Hausdorff operators on some product Hardy type spaces, Appl. Math. J. Chinese Univ. 27 (2012), 114–126.10.1007/s11766-012-2922-8Search in Google Scholar

[6] J. C. Chen and X. R. Zhu, Boundedness of multidimensional Hausdorff operators on H1(ℝn), J. Math. Anal. Appl. 409 (2014), 428–434.10.1016/j.jmaa.2013.07.042Search in Google Scholar

[7] V. I. Burenkov and E. Liflyand, On the boundedness of Hausdorff operators on Morrey-type spaces, Eurasian Math. J. 8 (2017), 97–104.Search in Google Scholar

[8] B. M. Damtew, Boundedness of multidimensional Hausdorff operator on Hardy-Morrey and Besov-Morrey spaces, J. Inequal. Appl. 2016 (2016), 293, 10.1186/s13660-016-1244-4.Search in Google Scholar

[9] K.-P. Ho, Hardy little woodpólya inequalities and Hausdorff operator on Block spaces, Math. Inequal. Appl. 19 (2016), 697–707.Search in Google Scholar

[10] C. Espinoza-Villalva and M. Guzmán-Partida, Continuity of Hardy type operators on rectangularly defined spaces, J. Math. Anal. Appl. 436 (2016), 29–38.10.1016/j.jmaa.2015.11.047Search in Google Scholar

[11] A. Hussain and G. Gao, Multidimensional Hausdorff operators and commutators on Herz-type spaces, J. Inequal. Appl. 2013 (2013), 594, 10.1186/1029-242X-2013-594.Search in Google Scholar

[12] G. P. Zhao and Q. Lou, Hausdorff operators on modulation spaces Mp,ps, J. Funct. Spaces 2018 (2018), 3048502, 10.1155/2018/3048502.Search in Google Scholar

[13] G. P. Zhao and W. C. Guo, Hausdorff operators on Sobolev spaces Wk,1, Integr. Transf. Special Funct. 30 (2019), 97–111.10.1080/10652469.2018.1537271Search in Google Scholar

[14] F. Gürbüz, Some estimates for generalized commutators of rough fractional maximal and integral operators on generalized weighted Morrey spaces, Canad. Math. Bull. 60 (2017), 131–145.10.4153/CMB-2016-067-8Search in Google Scholar

[15] F. Gürbüz, Multi-sublinear operators generated by multilinear fractional integral operators and commutators on the product generalized local Morrey spaces, Adv. Math. 47 (2018), 855–880.Search in Google Scholar

[16] F. Gürbüz, Generalized weighted Morrey estimates for Marcinkiewicz integrals with rough kernel associated with Schrödinger operator and their commutators, Chin. Ann. Math. Ser. B 41 (2020), 77–98.10.1007/s11401-019-0187-8Search in Google Scholar

[17] A. Hussain and A. Ajaib, Some results for the commutators of generalized Hausdorff operator, J. Math. Inequal. 13 (2019), 1129–1146.10.7153/jmi-2019-13-80Search in Google Scholar

[18] A. Hussain and G. Gao, Some new estimates for the commutatorsof n-dimensional Hausdorff operator, Appl. Math. J. Chinese Univ. 29 (2014), 139–150.10.1007/s11766-014-3169-3Search in Google Scholar

[19] X. M. Wu, Necessary and sufficient conditions for generalized Hausdorff operators and commutators, Anals Funct. Anal. 6 (2015), 60–72.10.15352/afa/06-3-6Search in Google Scholar

[20] A. Hussain and M. Ahmad, Weak and strong type estimates for the commutators of Huasdorff operators, Math. Inequal. Appl. 20 (2017), 49–56.Search in Google Scholar

[21] J. C. Chen, D. S. Fan, and S. Wang, Hausdorff operators on Euclidean spaces, Appl. Math. J. Chinese Univ. 28 (2013), 548–564.10.1007/s11766-013-3228-1Search in Google Scholar

[22] E. Lyflyand, Hausdorff operators on Hardy spaces, Eurasian Math. J. 4 (2013), 101–141.10.22405/2226-8383-2021-22-3-133-142Search in Google Scholar

[23] S. S. Volosivets, Multidimensional Hausdorff operator on p-adic fields, P-Adic Num. Ultrametric Anal. Appl. 2 (2010), 252–259.10.1134/S2070046610030076Search in Google Scholar

[24] S. S. Volosivets, Hausdorff Operators on p-adic linear spaces and their properties in Hardy, BMO, and Hölder spaces, Math. Notes 93 (2013), 382–391.10.1134/S0001434613030048Search in Google Scholar

[25] J. M. Ruan, D. S. Fan, and Q. Y. Wu, Weighted Herz space estimates for the Hausdorff operators on the Heisenberg group, Banach J. Math. Anal. 11 (2017), 513–535.10.1215/17358787-2017-0004Search in Google Scholar

[26] Q. Y. Wu and Z. W. Fu, Boundedness of Hausdorff operators on Hardy spaces in the Heisenberg group, Banach J. Math. Anal. 12 (2018), 909–934.10.1215/17358787-2018-0006Search in Google Scholar

[27] A. R. Mirotin, Boundedness of Hausdorff operators on real Hardy spaces H1 over locally compact groups, J. Inequal. Appl. 473 (2019), 519–533.10.1016/j.jmaa.2018.12.065Search in Google Scholar

[28] G. Folland and E. Stein, Hardy spaces on homogeneous groups, Math. Notes 28, Princeton Univ. Press, Princeton, 1982.10.1515/9780691222455Search in Google Scholar

[29] V. S. Guliyev, Two-weighted Lp-inequalities for singular integral operator on Heisenberg groups, Georgian Math. J. 1 (1994), 367–376.10.1515/GMJ.1994.367Search in Google Scholar

[30] T. Hytönen, C. Pérez, and E. Rela, Sharp reverse Hölder property for A∞ weights on spaces of homogeneous type, J. Funct. Anal. 263 (2012), 3883–3899.10.1016/j.jfa.2012.09.013Search in Google Scholar

[31] S. Indratno, D. Maldonado, and S. Silwal, A visual formalism for weights satisfying reverse inequalities, Expo. Math. 33 (2015), 1–29.10.1016/j.exmath.2013.12.008Search in Google Scholar

[32] B. Muckenhoupt, Weighted norm inequalities for the Hardy maximal function, Trans. Amer. Math. Soc. 165 (1972), 207–226.10.1090/S0002-9947-1972-0293384-6Search in Google Scholar

[33] J. García-Cuerva and J. Rubio de Francia, Weigted norm inequalities and related topics, North-Holland, Amsterdam, 1985.Search in Google Scholar

[34] M. J. Liu and S. Z. Lu, The continuity of some operators on Herz-type Hardy spaces on the Heisenberg group, Taiwanese J. Math. 16 (2012), 151–164.10.11650/twjm/1500406533Search in Google Scholar

[35] D. Yang, S. Lu, and G. Hu, Herz Type Spaces and Their Applications, Science Press, Beijing, 2008.Search in Google Scholar

[36] M. A. Ragusa, Homogeneous Herz spaces and regularity results, Nonlinear Anal. 71 (2009), e1909–e1914.10.1016/j.na.2009.02.075Search in Google Scholar

[37] D. S. Fan and D. Yang, Herz-type Hardy spaces on Vilenkin groups and their applications, Sci. China Ser. A 43 (2000), 481–494.10.1007/BF02897140Search in Google Scholar

[38] E. Hernández and D. Yang, Interpolation of Herz-type Hardy spaces, Illinois J. Math. 42 (1998), 564–581.10.1215/ijm/1255985461Search in Google Scholar

[39] S. Lu and D. Yang, Multiplier theorems for Herz type Hardy spaces, Proc. Am. Math. Soc. 126 (1998), 3337–3346.10.1090/S0002-9939-98-04931-4Search in Google Scholar

[40] L. Grafakos, X. Li, and D. Yang, Bilinear operators on Herz-type Hardy spaces, Trans. Amer. Math. Soc. 350 (1998), 1249–1275.10.1090/S0002-9947-98-01878-9Search in Google Scholar

[41] D. S. Fan and D. Yang, The weighted Herz-type Hardy spaces hK̇qα,p, Approx. Theory Appl. 13 (1997), 19–41.Search in Google Scholar

[42] A. Scapellato, Homogeneous Herz spaces with variable exponents and regularity results, Electron. J. Qual. Theory Differ. Equ. 82 (2018), 1–11, 10.14232/ejqtde.2018.1.82.Search in Google Scholar

[43] A. Scapellato, Regularity of solutions to elliptic equations on Herz spaces with variable exponents, Bound. Value Probl. 2019 (2019), 2, 10.1186/s13661-018-1116-6.Search in Google Scholar

[44] J. M. Ruan, D. S. Fan, and Q. Y. Wu, Weighted Morrey estimates for Hausdorff operator and its commutator on the Heisenberg group, Math. Inequal. Appl. 22 (2019), 307–329.Search in Google Scholar

[45] J. Alvares, M. Guzman-Partida, and J. Lakey, Spaces of bounded λ-central mean ocsillation, Morrey spaces, and λ-central Carleson measure, Collect. Math. 51 (2000), 1–47.Search in Google Scholar

[46] S. Z. Lu and D. C. Yang, The central BMO space and Little wood operators, Approx. Theory Appl. 11 (1995), 72–94.Search in Google Scholar

[47] J. M. Ruan and D. S. Fan, Hausdorff operators on the power weighted Hardy spaces, J. Math. Anal. Appl. 455 (2016), 31–48.10.1016/j.jmaa.2015.07.062Search in Google Scholar

[48] J. M. Ruan and D. S. Fan, Hausdorff operators on the weighted Herz-type Hardy spaces, Math. Inequal. Appl. 19 (2016), 565–587.Search in Google Scholar

[49] J. C. Chen, S. Y. He, and X. R. Zhu, Boundedness of Hausdorff operators on the power weighted Hardy spaces, Appl. Math. J. Chinese Univ. 32 (2017), 462–476.10.1007/s11766-017-3523-3Search in Google Scholar

[50] A. Hussain and A. Ajaib, Some weighted inequalities for Hausdorff operators and commutators, J. Inequal. Appl. 2018 (2018), 6, 10.1186/s13660-017-1588-4.Search in Google Scholar PubMed PubMed Central

[51] Q. X. Sun, D. S. Fan, and H. L. Li, Hausdorff operators on weighted Lorentz spaces, Korean J. Math. 2018, (2018), no. 26, 103–127.Search in Google Scholar

[52] A. Hussain and N. Sarfraz, The Hausdorff operator on weighted p-adic Morrey and Herz type spaces, P-Adic Num. Ultrametric Anal. Appl. 11 (2019), 151–162.10.1134/S2070046619020055Search in Google Scholar

[53] A. Hussain and N. Sarfraz, Estimates for the commutators of p-adic Hausdorff operator on Herz-Morrey spaces, Mathematics 7 (2019), 127, 10.3390/math7020127.Search in Google Scholar

Received: 2019-08-07

Revised: 2020-04-09

Accepted: 2020-04-28

Published Online: 2020-06-04

This work is licensed under the Creative Commons Attribution 4.0 International License.

Weighted CBMO estimates for commutators of matrix Hausdorff operator on the Heisenberg group

Abstract

1 Introduction

2 Some definitions and notations

Definition 2.1

Definition 2.2

Proposition 2.3

Proposition 2.4

Definition 2.5

Definition 2.6

Lemma 2.7

Proposition 2.8

3 Main results and their proofs

Theorem 3.1

Theorem 3.2

3.1 Proof of Theorem 3.1

3.2 Proof of Theorem 3.2

4 Conclusion

Acknowledgments

References

Journal and Issue

Articles in the same Issue