Home Reversed Hardy-Littlewood-Sobolev inequalities with weights on the Heisenberg group
Article Open Access

Reversed Hardy-Littlewood-Sobolev inequalities with weights on the Heisenberg group

  • Yunyun Hu EMAIL logo
Published/Copyright: November 20, 2023

Abstract

In this article, we establish some reverse weighted Hardy-Littlewood-Sobolev inequalities on the Heisenberg group. We then show the existence of extremal functions for the above inequalities by combining the subcritical approach and the renormalization method.

MSC 2010: 35A23; 45E10; 45G15; 43A85

1 Introduction

Stein and Weiss [34] established the following weighted Hardy-Littlewood-Sobolev (HLS) inequality:

(1.1) R n R n x α x y λ f ( x ) g ( y ) y β d x d y C n , α , β , p , q f L p ( R n ) g L q ( R n ) ,

where 1 < p , q < + , and α , β , and λ satisfy

1 p + 1 q + α + β + λ n = 2 , 1 p + 1 q 1 , α + β 0 , α < n p , β < n q , 0 < λ < n .

It is known as the Stein-Weiss inequality. A fundamental task in understanding Stein-Weiss inequality (1.1) is to study the sharp constant and existence of extremals. For q < p and α , β 0 , Lieb [30] proved the existence of extremals for the inequality (1.1) by using rearrangement inequalities. Moreover, he pointed out that the extremals cannot be expected to exist when p = q = 2 , λ = n 1 , α = 0 , and β = 1 , see also [27]. In the general case of p = q , Beckner [2,3] obtained the sharp constant of inequality (1.1).

The Stein-Weiss inequality (1.1) contains the following HLS inequality ( α = β = 0 ) established in [26,33]:

(1.2) R n R n x y λ f ( x ) g ( y ) d x d y C n , p , λ f L p ( R n ) g L q ( R n ) ,

where f L p ( R n ) , g L q ( R n ) , 1 < p , q < , 0 < λ < n , and 1 p + 1 q + λ n = 2 . Lieb [30] proved the existence of extremal functions for equation (1.2) by using rearrangement inequality. He also classified extremal functions and computed the sharp constant in the case of p = 2 , or q = 2 , or p = q = 2 n 2 n λ .

The HLS inequality was intensively studied since it plays, important role in many geometric problems by virtue of the complete knowledge of sharp constants and extremal functions. For example, the existence of extremal functions of HLS inequality (1.2) was considered on the Euclidean space [4,18,19,30], on Riemannian manifolds [24], and on the Heisenberg group [17].

In a ground-breaking work [20], Frank and Lieb obtained the sharp constant and the explicit form of extremal functions for Folland-Stein inequality on the Heisenberg group [17]. We recall some basic results concerning the Heisenberg group. The Heisenberg group H n C n × R is a Lie group with elements u = ( z , t ) and group multiplication

u v = ( z , t ) ( z , t ) = ( z + z , t + t + 2 Im ( z z ¯ ) ) ,

where z = ( z 1 , , z n ) , z = ( z 1 , , z n ) C n , t , t R , and z z ¯ = j = 1 n z j z j ¯ with z j = x j + 1 y j . The Lebesegue-Haar measure on H n is denoted by d u = d z d t . We define the homogeneous norm on H n by

u = ( z , t ) = ( z 4 + t 2 ) 1 4 ,

and the distance between u = ( z , t ) and v = ( z , t ) is

d ( u , v ) = v 1 u = ( z z 4 + ( t t 2 Im ( z z ¯ ) ) 2 ) 1 4 .

The natural dilation of the Heisenberg group is given by δ ( u ) = ( z , 2 t ) for any > 0 . With this norm, the ball centered at u H n with radius R > 0 is defined by B ( u , R ) = { v H n : v 1 u < R } . We also have the following triangle inequality:

(1.3) u v u + v u , v H n .

Throughout this article, the homogeneous dimension of H n is denoted by Q = 2 n + 2 . The Frank-Lieb inequality then states that for 0 < λ < Q and p = 2 Q 2 Q λ ,

(1.4) H n H n f ( u ) ¯ g ( v ) v 1 u λ d v d u 2 π n + 1 n ! λ Q n ! Γ Q λ 2 Γ 2 2 Q λ 4 f L p ( H n ) g L p ( H n )

with the equality if and only if

f ( u ) = c 1 g ( u ) = c 2 H ( δ ( a 1 u ) )

for some c 1 , c 2 C n , > 0 , and a H n (unless f 0 or g 0 ). Here, H is defined by

H ( z , t ) = ( ( 1 + z 2 ) 2 + t 2 ) 2 Q λ 4 .

More recently, Hang and Wang [25] gave a simpler proof of inequality (1.4), which bypasses the sophisticated proof for the existence of minimizers and Hersch-type arguments.

The u weighted HLS inequality (namely Stein-Weiss inequality) on the Heisenberg group was established by Han et al. [22]. Precisely, it states

(1.5) H n H n f ( u ) g ( v ) u α v 1 u λ v β d v d u C α , β , p , λ , n f L p ( H n ) g L q ( H n ) ,

where 1 < p , q < , 0 < λ < Q = 2 n + 2 , and α and β satisfy

α + β 0 , λ + α + β Q ,

α < Q p , β < Q q , 1 p + 1 q + α + β + λ Q = 2 .

Moreover, they proved the following z weighted HLS inequality on the Heisenberg group:

(1.6) H n H n f ( u ) g ( v ) z α v 1 u λ z β d v d u C α , β , p , λ , n f L p ( H n ) g L q ( H n )

for some constant C α , β , p , λ , n > 0 , where α and β are related via the following: α < 2 n p , β < 2 n q , 1 p + 1 q + α + β + λ Q = 2 for some 0 < λ < Q .

Recently, Han and Zhang [23] established the reverse version of inequality (1.4), i.e., for λ < 0 and p = 2 Q 2 Q λ , there exists a sharp constant N λ , Q such that

(1.7) H n H n f ( u ) ¯ g ( v ) v 1 u λ d v d u N λ , Q f L p ( H n ) g L p ( H n ) .

The sharp constant satisfies

( 8 B 1 ) λ Q 2 p 2 N λ , Q 2 π n + 1 n ! λ Q n ! Γ Q + λ 2 Γ 2 Q + 2 λ 2 ,

where B 1 B ( 0 , 1 ) and the volume of B 1 is defined as (see [12]):

B 1 = 2 π Q 2 2 Γ 1 2 Γ Q + 2 4 ( Q 2 ) Γ Q 2 2 Γ Q + 4 4 .

Different from the proof of [20], Han and Zhang proved the existence of extremal functions for inequality (1.7) by using the subcritical approach and renormalization method.

In the last two decades, various extensions of HLS inequality have been investigated. We focus on the typical examples of reverse HLS inequalities with weights. Chen et al. [6] established the following reverse Stein-Weiss inequality:

(1.8) R n R n x α x y λ f ( x ) g ( y ) y β d x d y C n , α , β , p , q f L p ( R n ) g L q ( R n ) ,

where p , q , α , β , and λ satisfy

1 p + 1 q + α + β + λ n = 2 , 0 < p , q < 1 ,

0 α < n p , 0 β < n q , 0 < λ < n .

They obtained the existence of extremal functions for inequality (1.8). Later, Chen et al. [5] established reverse Stein-Weiss inequalities on the upper half space and proved the existence of their extremal functions. Dai et al. [8] obtained the sharp reversed HLS inequality with an extended kernel on the upper half space, and computed the sharp constants. More recently, Dou et al. [13] proved the reversed HLS inequality with vertical weights on the upper half space and discussed the extremal functions by the renormalization method. For more results about (weighted) HLS-type inequalities, we refer the reader to [9,10,14,15,21,31,32] and the references therein.

In contrast, not much is known about reverse HLS-type inequalities on the Heisenberg group or the CR sphere. The main motivation of this article is to introduce the reverse HLS inequality with weights on the Heisenberg group. We first obtain the following reverse HLS-type inequality with z weights on H n .

Assume that α , β , λ , p , and q satisfy

(1.9) 0 < λ < Q = 2 n + 2 , 0 < p , q < 1 , 0 α < 2 n p , 0 β < 2 n q , 1 p + 1 q = λ + α + β Q + 2 ,

where 1 p + 1 p = 1 and 1 q + 1 q = 1 .

Theorem 1.1

Let α , β , λ , p , and q satisfy (1.9). Then, there exists a positive constant N α , β , λ N ( Q , α , β , λ , p , q ) such that for any nonnegative functions f L p ( H n ) and g L q ( H n ) ,

(1.10) H n H n z α v 1 u λ f ( u ) g ( v ) z β d v d u N α , β , λ f L p ( H n ) g L q ( H n )

holds. Moreover, the upper and lower bounds of constant N α , β , λ satisfy

(1.11) q ( p 1 ) 2 p q p q p 1 p p ( q 1 ) 2 p q p q q 1 q min { A 1 , A 2 } N α , β , λ min { A 1 , A 2 } ,

where

A 1 = ω n ( p 1 ) ( Q + α + λ ) p Q B 1 2 , ( 2 n + α ) p 2 n 4 ( p 1 ) p 1 p ω n ( q 1 ) ( Q + β ) q Q B 1 2 , ( 2 n + β ) q 2 n 4 ( q 1 ) q 1 q ,

A 2 = ω n ( p 1 ) ( Q + α ) p Q B 1 2 , ( 2 n + α ) p 2 n 4 ( p 1 ) p 1 p ω n ( q 1 ) ( Q + β + λ ) q Q B 1 2 , ( 2 n + β ) q 2 n 4 ( q 1 ) q 1 q ,

and ω n = 0 π d θ 1 0 π d θ 2 0 π d θ 2 n 2 0 2 π d θ 2 n 1 ( sin θ 1 ) 2 n 2 sin θ 2 n 2 is the 2 n -Lebesgue measure of the unitary Euclidean sphere in R 2 n .

Define operators

I α , β f ( v ) = H n z α v 1 u λ f ( u ) z β d u

and

E λ f ( v ) = H n v 1 u λ f ( u ) d u .

By duality and the reversed Hölder’s inequality, we readily check that the inequality (1.10) is equivalent to the following two inequalities.

Corollary 1.2

Let α , β , λ , p , and q satisfy equation (1.9). Then, there exists a constant C ( Q , α , β , λ , p ) > 0 depending on Q , α , β , λ , and p such that

(1.12) I α , β f L q ( H n ) C ( Q , α , β , λ , p ) f L p ( H n )

or

(1.13) ( E λ f ) z β L q ( H n ) C ( Q , α , β , λ , p ) z α f L p ( H n )

holds for any nonnegative function f L p ( H n ) .

We also obtain an improved form of inequality (1.10) without assumptions 0 α < 2 n p and 0 β < 2 n q .

Theorem 1.3

Let 0 < λ < Q = 2 n + 2 , 0 < p and q < 1 , and p , q satisfy 1 p + 1 q = λ + α + β Q + 2 . Then, there exists a positive constant S α , β , λ S ( Q , α , β , λ , p , q ) such that for any nonnegative functions f L p ( H n ) and g L q ( H n ) , we have

H n H n z α v 1 u λ f ( u ) g ( v ) z β d v d u S α , β , λ f L p ( H n ) g L q ( H n )

if one of the following conditions is satisfied:

  1. 0 α < 2 n p ,

  2. 0 β < 2 n q .

Moreover, for case ( i ) , S α , β , λ satisfies

q ( p 1 ) 2 p q p q p 1 p p ( q 1 ) 2 p q p q q 1 q A 2 S α , β , λ A 2 ,

and for case ( i i ) , S α , β , λ satisfies

q ( p 1 ) 2 p q p q p 1 p p ( q 1 ) 2 p q p q q 1 q A 1 S α , β , λ A 1 ,

where A 1 and A 2 are defined as in Theorem 1.1.

To prove Theorems 1.1 and 1.3, we adopt the standard approach, based on the reverse integral version of Hardy inequalities on the Heisenberg group (see [29]). The reverse integral version of Hardy inequalities on the upper half space was established by Chen et al. [5] and Tao [35]. Once we establish inequality (1.10), it is natural to consider the attainability of sharp constant N α , β , λ .

For the case p α = 2 Q 2 Q + 2 α + λ and q β = 2 Q 2 Q + 2 β + λ , we consider the following minimizing problem:

(1.14) N α , β , λ = inf H n H n z α v 1 u λ f ( u ) g ( v ) z β d v d u : f 0 , g 0 , f L p α ( H n ) = g L q β ( H n ) = 1 = inf H n H n z α v 1 u λ f ( u ) g ( v ) z β d v d u f L p α ( H n ) g L q β ( H n ) : f 0 , g 0 , f L p α ( H n ) , g L q β ( H n ) .

We obtain the following result for the extremal problem of inequality (1.10), which, to the best of our knowledge, is new.

Theorem 1.4

Let α , β , λ , p , and q satisfy equation (1.9). Then, N α , β , λ is attained by a pair of positive functions ( f , g ) L p α ( H n ) × L q β ( H n ) satisfying f L p α ( H n ) = 1 and g L q β ( H n ) = 1 .

Remark 1.5

Note that for p = p α = 2 Q 2 Q + 2 α + λ and q = q β = 2 Q 2 Q + 2 β + λ , using Cayley transform inequality (1.10) is equivalent to the following inequality on sphere S 2 n + 1 (see details in Section 3).

(1.15) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η C α , β , λ F L p α ( S 2 n + 1 ) G L q β ( S 2 n + 1 ) ,

where C α , β , λ is a positive constant depending on Q , α , β , and λ . Hence, the existence of extremal functions for inequality (1.10) is in fact the existence problem of equation (1.15).

For α = β = Q λ 2 and p = q = 2 , Beckner [1] obtained the sharp constant of inequality (1.6) and the nonexistence of optimizers. Recently, in general case q < p , Chen et al. [7] proved the existence of extremal functions for inequality (1.5) under assumption α + β 0 . This result is more general than Lieb’s results (see [30]) in the sense that they removed the restriction of α , β 0 . Lieb [30] used the rearrangement inequality to prove the existence of extremals for inequality (1.1) in Euclidean spaces. Due to the absence of rearrangement inequality in the Heisenberg group, Chen et al. [7] obtained the existence of the maximizers for inequalities (1.5) and (1.6) by employing the concentration compactness principle.

In this article, we focus on the less investigated inequality (1.10) and the existence of extremal functions. We verify the existence of minimizers for inequality (1.10) by combining the subcritical approach (see [10,16]) and the renormalization method (see [13,23,36]). We cannot simply follow the same line of [13,16,23,36] to prove the existence of extremal functions for inequality (1.10) due to the loss of translation invariance of reverse weighted HLS inequality and the complex structure of the Heisenberg group. To overcome these difficulties, we use the subcritical approach to construct the extremal sequences of inequality (1.15), then combine the renormalization method and rotation invariance to obtain the expected attainability of the best constant. We believe this method can also be applied to obtain the existence of extremal function for several other weighted integral inequalities on the Heisenberg group.

We also investigate the reverse Stein-Weiss inequality in the Heisenberg group.

Theorem 1.6

Let 0 < λ < Q = 2 n + 2 , 0 < p , q < 1 , and 1 p + 1 q = λ + α + β Q + 2 . Then, there exists a positive constant N Q , α , β , λ , p > 0 such that for any nonnegative functions f L p ( H n ) and g L q ( H n ) , we have

(1.16) H n H n u α v 1 u λ f ( u ) g ( v ) v β d v d u N Q , α , β , λ , p f L p ( H n ) g L q ( H n )

if one of the following conditions is satisfied:

  1. 0 α < Q p ,

  2. 0 β < Q q .

Moreover, for case ( i ) , N Q , α , β , λ , p satisfies

q ( p 1 ) 2 p q p q p 1 p p ( q 1 ) 2 p q p q q 1 q A 2 N Q , α , β , λ , p A 2 ,

and for case ( i i ) , N Q , α , β , λ , p satisfies

q ( p 1 ) 2 p q p q p 1 p p ( q 1 ) 2 p q p q q 1 q A 1 N Q , α , β , λ , p A 1 ,

where

A 1 = ω n ( p 1 ) ( Q + α + λ ) p Q B 1 2 , n + 1 2 p 1 p ω n ( q 1 ) ( Q + β ) q Q B 1 2 , n + 1 2 q 1 q ,

A 2 = ω n ( p 1 ) ( Q + α ) p Q B 1 2 , n + 1 2 p 1 p ω n ( q 1 ) ( Q + β + λ ) q Q B 1 2 , n + 1 2 q 1 q ,

where ω n is defined as Theorem 1.1.

Note that for Q = 2 n + 2 , the reverse Stein-Weiss inequality on the homogeneous group G established by Kassymov et al. [28] also contains the case of Heisenberg group. Different from the study by Kassymov et al. [28], we obtain the upper and lower bounds of constant N Q , α , β , λ , p for inequality (1.16).

Similar to Theorem 1.1, we can also prove Theorem 1.6. Hence, we omit the details of the proof of Theorem 1.6.

Define operators

R α , β f ( v ) = H n u α v 1 u λ f ( u ) v β d u .

It is easy to verify that the inequality (1.16) is equivalent to the following two inequalities.

Corollary 1.7

Under the hypothesis of Theorem 1.6, there exists a constant C ( Q , α , β , λ , p ) > 0 depending on Q , α , β , λ , and p such that

R α , β f L q ( H n ) C ( Q , α , β , λ , p ) f L p ( H n )

or

( E λ f ) v β L q ( H n ) C ( Q , α , β , λ , p ) u α f L p ( H n )

holds for any nonnegative function f L p ( H n ) .

This article is organized as follows: Section 2 is devoted to establishing the reverse weighted HLS inequality (1.10); in Section 3, combining the subcritical approach with the renormalization method, we prove the existence of extremal function of the reverse weighted HLS inequality (1.10).

Throughout the article, positive constants are denoted by c and C (with subscript in some cases) and are allowed to vary within a single line or formula.

2 The reverse weighted HLS inequality on the Heisenberg group

In this section, we prove the reverse weighted HLS inequalities on the Heisenberg group. The following lemma is the reverse integral version of Hardy inequalities on general homogeneous groups in the study by Kassymov et al. [29], which will be crucial in our proof of Theorem 1.2.

Lemma 2.1

[29] Let G be a homogeneous group of homogeneous dimension Q. Assume that q ( 0 , 1 ) and p < 0 . Suppose that W , U 0 are locally integrable functions on G . Then, the inequality

(2.1) G B ( 0 , u ) f ( v ) d v p W ( u ) d u 1 p C 1 ( p , q ) G f q ( u ) U ( u ) d u 1 q

holds for some C 1 ( p , q ) > 0 and all nonnegative measurable functions f if and only if

(2.2) 0 < A 1 = inf u G G \ B ( 0 , u ) W ( v ) d v 1 p B ( 0 , u ) U 1 q ( v ) d v 1 q .

Moreover, the biggest constant C 1 ( p , q ) > 0 in equation (2.1) has the following relation to A 1 :

A 1 C 1 ( p , q ) q q + p 1 p p p + q 1 q A 1 .

In addition, inequality

(2.3) G G \ B ( 0 , u ) f ( v ) d v p W ( u ) d u 1 p C 2 ( p , q ) G f q ( u ) U ( u ) d u 1 q

holds for some C 2 ( p , q ) > 0 and all nonnegative measurable functions f if and only if

(2.4) 0 < A 2 = inf u G B ( 0 , u ) W ( v ) d v 1 p G \ B ( 0 , u ) U 1 q ( v ) d v 1 q .

Moreover, the biggest constant C 2 ( p , q ) > 0 in equation (2.3) has the following relation to A 2 :

A 2 C 2 ( p , q ) q q + p 1 p p p + q 1 q A 2 .

Remark 2.2

Note that for Q = 2 n + 2 , the homogeneous group G contains the Heisenberg group H n . Hence, Lemma 2.1 also holds on the Heisenberg group H n .

Proof of Theorem 1.1

By reversed Hölder’s inequality, we have

H n H n z α v 1 u λ f ( u ) g ( v ) z β d v d u = H n H n z α v 1 u λ g ( v ) z β d v f ( u ) d u H n H n z α v 1 u λ g ( v ) z β d v q d u 1 q f L p ( H n ) .

Let h ( v ) = z β g ( v ) . It is easy to see that equation (1.12) is equivalent to

H n H n z α v 1 u λ h ( v ) d v p d u N α , β , λ p z β h L q ( H n ) p .

Since

H n z α v 1 u λ h ( v ) d v B 0 , u 2 z α v 1 u λ h ( v ) d v ,

we have

H n z α v 1 u λ h ( v ) d v p B 0 , u 2 z α v 1 u λ h ( v ) d v p .

Therefore, we obtain

(2.5) H n z α p H n v 1 u λ h ( v ) d v p d u 1 p H n z α p B 0 , u 2 v 1 u λ h ( v ) d v p d u 1 p I 1 1 p ( u ) .

Similarly, one has

(2.6) H n z α p ( H n v 1 u λ h ( v ) d v ) p d u 1 p H n z α p H n \ B ( 0 , 2 u ) v 1 u λ h ( v ) d v p d u 1 p I 2 1 p ( u ) .

Combining equations (2.5) and (2.6), we have

H n z α p H n v 1 u λ h ( v ) d v p d u 1 p I 1 1 p ( u ) 2 + I 2 1 p ( u ) 2 .

Hence, it suffices to show that

I i ( u ) N α , β , λ p z β h L q ( H n ) p , i = 1 , 2 .

We divide into two cases to discuss.

(1) If v u 2 , then we have

u = u 1 = u 1 v v 1 u 1 v + v 1 v 1 u + u 2 .

Then, for any λ > 0 , we obtain

2 λ u λ v 1 u λ .

This implies that

2 λ B 0 , u 2 u λ h ( v ) d v B 0 , u 2 v 1 u λ h ( v ) d v ,

so that

B 0 , u 2 v 1 u λ h ( v ) d v p 2 λ p B 0 , u 2 u λ h ( v ) d v p .

Therefore, we have

(2.7) I 1 ( u ) = H n z α p B 0 , u 2 v 1 u λ h ( v ) d v p d u 2 λ p H n z α p B 0 , u 2 u λ h ( v ) d v p d u .

Set W ( u ) = z α p u λ p and U ( v ) = z β q in equation (2.1), it follows from equation (2.7) that

(2.8) I 1 ( u ) 2 λ p N α , β , λ p z β h L q ( H n ) p ,

where N α , β , λ can be seen as C 1 ( p , q ) in equation (2.1). Now, we verify the condition (2.2). Since β < 2 n q , we have

1 p + 1 q = λ + α + β Q + 2 < λ + α Q 1 q + 2 .

That is, Q + ( α + λ ) p < 0 . Using polar decomposition, we have

(2.9) H n \ B ( 0 , u ) W ( v ) d v = H n \ B ( 0 , u ) z α p v λ p d v = ω n u ρ ( α + λ ) p + Q 1 d ρ 0 π ( sin θ ) α p 2 + n 1 d θ = ω n Q + ( α + λ ) p B 1 2 , 2 n + α p 4 u Q + ( α + λ ) p ,

where ω n = 0 π d θ 1 0 π d θ 2 0 π d θ 2 n 2 0 2 π d θ 2 n 1 ( sin θ 1 ) 2 n 2 sin θ 2 n 2 is the 2 n -Lebesgue measure of the unitary Euclidean sphere in R 2 n .

Moreover, since β < 2 n q , we have

β q ( 1 q ) + Q > β q ( 1 q ) β q = 0 .

Then, we obtain

(2.10) B ( 0 , u ) U 1 q ( v ) d v = B ( 0 , u ) z β q ( 1 q ) d v = ω n 0 u ρ Q 1 β q ( 1 q ) d ρ 0 π ( sin θ ) β q ( 1 q ) 2 + n 1 d θ = ω n Q + β q B 1 2 , 2 n + β q 4 u Q + β q .

Combining equations (2.9) and (2.10) yields

A 1 = inf u H n H n \ B ( 0 , u ) W ( v ) d v 1 p B ( 0 , u ) U 1 q ( v ) d v 1 q = ω n Q + ( α + λ ) p B 1 2 , 2 n + α p 4 1 p ω n Q + β q B 1 2 , 2 n + β q 4 1 q inf u H n u Q + ( α + λ ) p p + Q + β q q = ω n Q + ( α + λ ) p B 1 2 , 2 n + α p 4 1 p ω n Q + β q B 1 2 , 2 n + β q 4 1 q > 0 ,

where in the second equality we have used the fact Q + ( α + λ ) p p + Q + β q q = 0 . Therefore, we prove equation (2.8).

(2) If 2 u v , then v 2 v 1 u . Similar to the case I 1 ( u ) , let W ( u ) = z α p and U ( v ) = z β q v λ q in equation (2.3), we have

(2.11) I 2 ( u ) = H n z α p H n \ B ( 0 , 2 u ) v 1 u λ h ( v ) d v p d u 2 λ p H n z α p H n \ B ( 0 , 2 u ) v λ h ( v ) d v p d u 2 λ p N α , β , λ p z β h L q ( H n ) p ,

where N α , β , λ can be seen as C 2 ( p , q ) in equation (2.3).

Let us verify the condition (2.4). Using equation (1.9), we have Q + ( β + λ ) q < 0 . Then, a simple computation gives

(2.12) B ( 0 , u ) W ( u ) d u = B ( 0 , u ) z α p d u = ω n 0 u ρ α p + Q 1 d ρ 0 π ( sin θ ) α p 2 + n 1 d θ = ω n Q + α p B 1 2 , 2 n + α p 4 u Q + α p

and

(2.13) H n \ B ( 0 , u ) U 1 q ( v ) d v = H n \ B ( 0 , u ) ( z β q v λ q ) 1 q d v = ω n u ρ ( β + λ ) q + Q 1 d ρ 0 π ( sin θ ) β q 2 + n 1 d θ = ω n Q + ( β + λ ) q B 1 2 , 2 n + β q 4 u Q + ( β + λ ) q .

Combining equations (2.12) and (2.13), we know that

A 2 = ω n Q + α p B 1 2 , 2 n + α p 4 1 p ω n Q + ( β + λ ) q B 1 2 , 2 n + β q 4 1 q ,

and hence, equation (2.11) is proved.

Now we give the estimate of constant N α , β , λ . From Lemma 2.1, we know

q q + p 1 p p p + q 1 q min { A 1 , A 2 } N α , β , λ min { A 1 , A 2 } ,

where

A 1 = ω n Q + ( α + λ ) p B 1 2 , 2 n + α p 4 1 p ω n Q + β q B 1 2 , 2 n + β q 4 1 q ,

A 2 = ω n Q + α p B 1 2 , 2 n + α p 4 1 p ω n Q + ( β + λ ) q B 1 2 , 2 n + β q 4 1 q .

This completes the proof of Theorem 1.1.□

Proof of Theorem 1.3

From the proof process of Theorem 1.1, we immediately prove Theorem 1.3.

3 Existence of extremal functions for reverse HLS inequality with horizontal weights

This section is devoted to proving the existence of extremal functions for inequality (1.10) by combining the subcritical approach and the renormalization method.

To better illustrate our results, we first introduce some well-known notations. The sphere S 2 n + 1 is the boundary of unit ball in C n + 1 . Define S 2 n + 1 = { ξ = ( ξ 1 , , ξ n + 1 ) C n : ξ ξ ¯ = j = 1 n + 1 ξ j 2 = 1 } . On the sphere, the distance function is defined as follows:

d 2 ( ξ , η ) = 2 1 ξ η ¯ = ξ η 2 2 i Im ( ξ η ¯ ) = ( ξ η 4 + 4 Im 2 ( ξ η ¯ ) ) 1 2 .

The standard Euclidean volume element of S 2 n + 1 will be denoted by d ξ .

The Heisenberg group H n is conformally equivalent to S 2 n + 1 via the Cayley transform C : H n S 2 n + 1 \ { 0 , 0 , , 0 , 1 } given by

C ( z , t ) = 2 z 1 + z 2 + i t , 1 z 2 i t 1 + z 2 + i t

and with the inverse

C 1 ( ξ ) = ξ 1 1 + ξ n + 1 , , ξ n 1 + ξ n + 1 , Im 1 ξ n + 1 1 + ξ n + 1 .

The Jacobian determinant of this transformation is given by

J C ( z , t ) = 2 2 n + 1 ( ( 1 + z 2 ) 2 + t 2 ) n + 1 ,

which implies that

S 2 n + 1 ϕ ( ξ ) d ξ = H n ( ϕ C ) J C d u .

Under Cayley transform, a simple computation gives

1 ξ η ¯ = 2 v 1 u 2 ( ( 1 + z 2 ) 2 + t 2 ) 1 2 ( ( 1 + z 2 ) 2 + ( t ) 2 ) 1 2 ,

where ξ = C ( u ) and η = C ( v ) . That is,

d ( ξ , η ) = d ( u , v ) 4 ( 1 + z 2 ) 2 + t 2 1 4 4 ( 1 + z 2 ) 2 + ( t ) 2 1 4 .

For any F L p ( S 2 n + 1 ) and G L q ( S 2 n + 1 ) , there exist corresponding functions

f ( u ) = J C ( u ) 1 p F ( C ( u ) ) L p ( H n ) and g ( v ) = J C ( v ) 1 q G ( C ( v ) ) L q ( H n )

such that

F L p ( S 2 n + 1 ) = f L p ( H n ) and G L q ( S 2 n + 1 ) = g L q ( H n ) ,

respectively.

For p α = 2 Q 2 Q + 2 α + λ and q β = 2 Q 2 Q + 2 β + λ , it is easy to verify that the inequality (1.10) is equivalent to the following inequality on sphere S 2 n + 1 :

(3.1) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η C α , β , λ F L p α ( S 2 n + 1 ) G L q β ( S 2 n + 1 ) ,

where C α , β , λ is a positive constant depending on Q , α , β , and λ .

We define the extremal function of inequality (3.1) as

C α , β , λ = inf F L p α ( S 2 n + 1 ) = 1 G L q β ( S 2 n + 1 ) = 1 S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η = inf F L p α ( S 2 n + 1 ) , G L q β ( S 2 n + 1 ) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F L p α ( S 2 n + 1 ) G L q β ( S 2 n + 1 ) .

Then, we have

Theorem 3.1

C α , β , λ is attained by a pair of positive functions ( F , G ) L p α ( S 2 n + 1 ) × L q β ( S 2 n + 1 ) satisfying F L p α ( S 2 n + 1 ) = 1 and G L q β ( S 2 n + 1 ) = 1 , respectively.

3.1 Subcritical reverse weighted HLS inequality on CR sphere

The aim of this subsection is to establish the subcritical reverse weighted HLS inequality on S 2 n + 1 . We also prove the existence of the corresponding extremal functions, which is inspired by [10,13,16].

Theorem 3.2

Assume 0 < λ < Q , p ( 0 , p α ) , and q ( 0 , q β ) . Then, there exists a constant C Q , α , β , λ , p > 0 depending on Q , α , β , λ , and p such that for any nonnegative functions F L p ( S 2 n + 1 ) and G L q ( S 2 n + 1 ) , there holds

(3.2) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η C Q , α , β , λ , p F L p ( S 2 n + 1 ) G L q ( S 2 n + 1 ) .

For η B n + 1 , we introduce the operator

T λ , α , β F ( η ) = S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 F ( ξ ) d ξ .

We can easily see that the inequality (3.2) is equivalent to

(3.3) T λ , α , β F L q ( S 2 n + 1 ) C Q , α , β , λ , p F L p ( S 2 n + 1 )

for any nonnegative function F L p ( S 2 n + 1 ) .

Now we consider the extremal problem of inequality (3.2):

(3.4) C Q , α , β , λ , p * = inf S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η : F 0 , G 0 , F L p ( S 2 n + 1 ) = G L q ( S 2 n + 1 ) = 1 = inf S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F L p ( S 2 n + 1 ) G L q ( S 2 n + 1 ) : F 0 , G 0 , F L p ( S 2 n + 1 ) , G L q ( S 2 n + 1 ) .

By equations (3.2) and (3.4), we know that C Q , α , β , λ , p * C Q , α , β , λ , p > 0 .

The following proposition will be used in the proof of Theorem 1.4.

Proposition 3.3

( i ) There exists a pair of nonnegative functions ( F , G ) L 1 ( S 2 n + 1 ) × L 1 ( S 2 n + 1 ) such that F L p ( S 2 n + 1 ) = G L q ( S 2 n + 1 ) = 1 and

C Q , α , β , λ , p * = S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η .

( i i ) For ξ S 2 n + 1 , the functions F and G satisfy the Euler-Lagrange equations

(3.5) C Q , α , β , λ , p * F p 1 ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η , C Q , α , β , λ , p * G q 1 ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) β 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) α 2 F ( η ) d η .

( i i i ) There exists some constant C C ( Q , α , β , λ , p , q ) > 0 such that

(3.6) 1 C F , G C .

Furthermore, F , G C γ ( S 2 n + 1 ) with γ ( 0 , 1 ) .

Proof

(i). We first show that C n , α , β , λ , p * is attained by a pair of nonnegative functions ( F , G ) L 1 ( S 2 n + 1 ) × L 1 ( S 2 n + 1 ) .

We choose a pair of nonnegative minimizing sequence { F j , G j } j = 1 + C ( S 2 n + 1 ) × C ( S 2 n + 1 ) such that

F j L p ( S 2 n + 1 ) = G j L q ( S 2 n + 1 ) = 1

and

C Q , α , β , λ , p * = lim j + S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η .

We divide the proof of part (i) into three steps.

Step 1. Our first step is to claim that

(3.7) F j L 1 ( S 2 n + 1 ) C 1 , G j L 1 ( S 2 n + 1 ) C 2 uniformly .

Indeed, we know from equation (1.11) that there exist constants C 3 > 0 and C 4 > 0 such that

0 < C 3 S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η C 4 < .

Then, using reversed Hölder’s inequality, we obtain

T λ , α , β F j L q ( S 2 n + 1 ) = T λ , α , β F j L q ( S 2 n + 1 ) G j L q ( S 2 n + 1 ) C 4 ,

T λ , α , β G j L p ( S 2 n + 1 ) = T λ , α , β G j L p ( S 2 n + 1 ) F j L p ( S 2 n + 1 ) C 4 .

Since q β < q < 0 , we know that for some constant M > 0 (to be determined later),

(3.8) C 4 q S 2 n + 1 T λ , α , β F j ( η ) q d η = { T λ , α , β F j ( η ) M } T λ , α , β F j ( η ) q d η + { T λ , α , β F j ( η ) < M } T λ , α , β F j ( η ) q d η M q S 2 n + 1 + { T λ , α , β F j ( η ) < M } 1 q q β T λ , α , β F j L q β ( S 2 n + 1 ) q ,

where 1 q β + 1 q β = 1 . By equation (3.3) and reversed Hölder’s inequality, we deduce

(3.9) T λ , α , β F j L q β ( S 2 n + 1 ) C 5 F j L p α ( S 2 n + 1 ) C 5 S 2 n + 1 1 p α 1 p F j L p ( S 2 n + 1 ) = C 5 S 2 n + 1 1 p α 1 p .

We now choose M > 0 such that M q S 2 n + 1 = C 4 q 2 . Using equations (3.8) and (3.9), we have

C 4 q 2 C 5 S 2 n + 1 1 p α 1 p q { T λ , α , β F j ( η ) < M } 1 q q β ,

which helps to conclude that

{ T λ , α , β F j ( η ) < M } C 4 2 C 5 S 2 n + 1 1 p α 1 p q β q q β q > 0 .

Hence, there exists ε 0 > 0 such that for any j , we can find two points η j 1 , η j 2 Ω 1 = { η : T λ , α , β F j ( η ) < M } satisfying η j 1 η j 2 ε 0 . Then, we have

S 2 n + 1 F j ( ξ ) d ξ C 6 S 2 n + 1 \ B ( η j 1 , ε 0 4 ) F j ( ξ ) d ξ + C 6 S 2 n + 1 \ B ( η j 2 , ε 0 4 ) F j ( ξ ) d ξ C ( ε 0 ) S 2 n + 1 \ B ( η j 1 , ε 0 4 ) ( 1 ξ n + 1 2 ) α 2 1 ξ η j 1 ¯ λ 2 ( 1 η n + 1 1 2 ) β 2 F j ( ξ ) d ξ + C ( ε 0 ) S 2 n + 1 \ B ( η j 2 , ε 0 4 ) ( 1 ξ n + 1 2 ) α 2 1 ξ η j 2 ¯ λ 2 ( 1 η n + 1 2 2 ) β 2 F j ( ξ ) d ξ 2 C ( ε 0 ) M uniformly for all j .

From this, we know that F j L 1 ( S 2 n + 1 ) C 1 . A similar argument also shows that G j L 1 ( S 2 n + 1 ) C 2 .

Step 2. There exist two subsequences of { F j p } and { G j p } (still denoted by { F j p } and { G j p } ) and two nonnegative functions F , G L 1 ( S 2 n + 1 ) such that

(3.10) S 2 n + 1 F j ( ξ ) p d ξ S 2 n + 1 F ( ξ ) p d ξ , S 2 n + 1 G j ( ξ ) q d ξ S 2 n + 1 G ( ξ ) q d ξ as j + .

Without loss of generality, we may assume that p q . By equation (3.7), we infer that there exist two subsequences of { F j p } and { G j p } (still denoted by { F j p } and { G j p } ) and two nonnegative functions F , G L 1 ( S 2 n + 1 ) such that

F j p F p weakly in L 1 p ( S 2 n + 1 ) ,

G j p G p weakly in L 1 p ( S 2 n + 1 ) .

Since 1 L 1 1 p ( S 2 n + 1 ) and G j q p L 1 q p ( S 2 n + 1 ) L 1 1 p ( S 2 n + 1 ) , we immediately derive equation (3.10).

Step 3: We show that

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η as j .

By equation (3.7) and the interpolation inequality, we obtain

S 2 n + 1 F j ( ξ ) p d ξ C > 0 and S 2 n + 1 G j ( ξ ) q d ξ C > 0 .

It follows that

S 2 n + 1 F ( ξ ) p d ξ C > 0 and S 2 n + 1 G ( ξ ) q d ξ C > 0 .

Then, for any fixed η S 2 n + 1 ,

( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 F 1 p ( ξ ) L 1 1 p ( S 2 n + 1 ) ,

and for any fixed ξ S 2 n + 1 ,

( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G 1 p ( η ) L 1 1 p ( S 2 n + 1 ) .

Thus, as j ,

S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 F j p ( ξ ) F 1 p ( ξ ) d ξ S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 F ( ξ ) d ξ , S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G j p ( η ) G 1 p ( η ) d η S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η .

Similar to the arguments in the study by Dou et al. [11], we know that the above convergences are uniformly convergent for all ξ , η S 2 n + 1 .

Therefore, for any ε > 0 small enough, there exists j 0 N such that for all j > j 0 ,

(3.11) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j p ( ξ ) F 1 p ( ξ ) G j p ( η ) G 1 p ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j p ( ξ ) F 1 p ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η ε S 2 n + 1 F j p ( ξ ) F 1 p ( ξ ) d ξ C ε ,

where in the last inequality, we have used Hölder’s inequality. Noting that f 1 p L 1 1 p ( S 2 n + 1 ) and

S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η C S 2 n + 1 G ( η ) d η C ,

we arrive at

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j p ( ξ ) F 1 p ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η as j .

From the above convergence and equation (3.11), we obtain

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j p ( ξ ) F 1 p ( ξ ) G j p ( η ) G 1 p ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η as j .

Again by Hölder’s inequality, we have

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j p ( ξ ) F 1 p ( ξ ) G j p ( η ) G 1 p ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η p × S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η 1 p .

We can then readily check that

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η liminf j + S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η .

Combining the preceding estimate with equation (3.4) gives

C Q , α , β , λ , p * = S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η = lim j + S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F j ( ξ ) G j ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η .

Therefore, we deduce that ( F , G ) L 1 ( S 2 n + 1 ) × L 1 ( S 2 n + 1 ) is a minimizer.

(ii). We show that F and G satisfy Euler-Lagrange equation (3.5).

To see this, we first prove F > 0 and G > 0 a.e. on S 2 n + 1 . For any positive function φ C ( S 2 n + 1 ) and sufficiently small t > 0 , we have F + t φ > 0 on S 2 n + 1 and

(3.12) t S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 φ ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η = S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( F + t φ ) ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η C Q , α , β , λ , p * ( F + t φ L p ( S 2 n + 1 ) F L p ( S 2 n + 1 ) ) = C Q , α , β , λ , p * t S 2 n + 1 ( F + θ φ ) p ( ξ ) d ξ 1 p 1 S 2 n + 1 ( F + θ φ ) p 1 ( ξ ) φ ( ξ ) d ξ ( 0 < θ < t ) ,

where in the last equality, we have used the mean value theorem. Then, applying equation (3.12) and Fatou’s lemma, we have

S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 φ ( ξ ) G ( η ) d ξ d η C Q , α , β , λ , p * lim t 0 + S 2 n + 1 ( F + θ φ ) p ( ξ ) d ξ 1 p 1 S 2 n + 1 ( F + θ φ ) p 1 ( ξ ) φ ( ξ ) d ξ C Q , α , β , λ , p * S 2 n + 1 F p 1 ( ξ ) φ ( ξ ) d ξ .

Now we claim F ( ξ ) > 0 a.e. on S 2 n + 1 . If not, then for any ε > 0 , there exists Ω ε S 2 n + 1 such that Ω ε > 0 and

F ( ξ ) < ε ξ Ω ε .

Then, using equation (3.12), we derive

ε q 1 Ω ε Ω ε F p 1 ( ξ ) d ξ 1 C Q , α , β , λ , p * S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d ξ d η C B n + 1 G ( η ) d η C .

This gives us a contradiction when ε > 0 is sufficiently small. Again by repeating the whole procedure above for G , we have G > 0 a.e. on S 2 n + 1 . Therefore, ( F , G ) is a pair of solutions of (3.5).

(iii). We prove ( F , G ) C γ ( S 2 n + 1 ) × C γ ( S 2 n + 1 ) with γ ( 0 , 1 ) .

Since F L 1 ( S 2 n + 1 ) and 0 < p < p α < 1 , using equation (3.5), there exists a constant C 7 > 0 such that 1 C 7 < F < C 7 . Similarly, for some constant C 8 > 0 , we have 1 C 8 < G < C 8 .

Let W ( ξ ) = C Q , α , β , λ , p * F p 1 ( ξ ) . For any given ξ 1 , ξ 2 S 2 n + 1 and arbitrary η S 2 n + 1 , a simple computation gives

(3.13) 1 ξ 1 η ¯ λ 2 1 ξ 2 η ¯ λ 2 C ξ 1 ξ 2 λ 2 , if 0 < λ < 2 , C ξ 1 ξ 2 , if λ 2 .

Using the boundedness of G and equation (3.13), we conclude that for any given ξ 1 , ξ 2 S 2 n + 1 and α ( 0 , 1 ) ,

(3.14) W ( ξ 1 ) W ( ξ 2 ) = S 2 n + 1 ( 1 ξ n + 1 1 2 ) α 2 1 ξ 1 η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η S 2 n + 1 ( 1 ξ n + 1 2 2 ) α 2 1 ξ 2 η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η = S 2 n + 1 ( 1 ξ n + 1 1 2 ) α 2 ( 1 ξ n + 1 2 2 ) α 2 1 ξ 1 η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η S 2 n + 1 ( 1 ξ n + 1 2 2 ) α 2 ( 1 ξ 1 η ¯ λ 2 1 ξ 2 η ¯ λ 2 ) ( 1 η n + 1 2 ) β 2 G ( η ) d η C ξ n + 1 1 ξ n + 1 2 α 2 S 2 n + 1 G ( η ) d η + C ξ 1 ξ 2 γ S 2 n + 1 G ( η ) d η C ξ 1 ξ 2 γ ,

where γ 0 , α 2 . Moreover, for α = 0 and α 2 , one can easily see that W ( ξ 1 ) W ( ξ 2 ) C ξ 1 ξ 2 γ . This implies that W is Hölder continuous on S 2 n + 1 , and hence F is at least Hölder continuous on S 2 n + 1 . Repeating the above argument shows that G is at least Hölder continuous on S 2 n + 1 .□

3.2 Minimizer of reverse weighted HLS inequality on CR sphere

In this subsection, we prove Theorem 3.1. The method we shall use is similar to that used by Dou et al. [13], but we have to set up a framework to fit the complex structure of Heisenberg group. We first give the following lemma, which is crucial in our proof.

Lemma 3.4

Let { F p , G q } C γ ( S 2 n + 1 ) × C γ ( S 2 n + 1 ) be defined as in Proposition 3.3. Then, there holds

C Q , α , β , λ , p * N α , β , λ a s p p α , q q β ,

and the minimizer pair { F p , G q } satisfies

(3.15) N α , β , λ = lim p p α q q β S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F p ( ξ ) G q ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F p L p α ( S 2 n + 1 ) G q L q β ( S 2 n + 1 ) .

Proof

Set F ˜ p = F p F p L p α ( S 2 n + 1 ) , G ˜ q = G q G q L q β ( S 2 n + 1 ) , and noting that F p L p ( S 2 n + 1 ) = G q L q ( S 2 n + 1 ) = 1 , and by reversed Hölder’s inequality, we find

C Q , α , β , λ , p * = F p L p α ( S 2 n + 1 ) G q L q β ( S 2 n + 1 ) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ˜ p ( ξ ) G ˜ q ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 1 q β 1 q + 1 p α 1 p S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ˜ p ( ξ ) G ˜ q ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η N α , β , λ as p p α , q q β .

This implies

(3.16) liminf p p α , q q β C Q , α , β , λ , p * N α , β , λ .

Let { F k , G k } L p α ( S 2 n + 1 ) × L q β ( S 2 n + 1 ) be a pair of minimizing sequence of N α , β , λ . It follows

N α , β , λ = lim k + S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F k ( ξ ) G k ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F k L p α ( S 2 n + 1 ) G k L q β ( S 2 n + 1 ) .

Define F ˜ k = F k F k L p ( S 2 n + 1 ) and G ˜ k = G k G k L q ( S 2 n + 1 ) . For p ( 0 , p α ) and q ( 0 , q β ) , we have

(3.17) C Q , α , β , λ , p * S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F ˜ k ( ξ ) G ˜ k ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F ˜ k L p ( S 2 n + 1 ) G ˜ k L q ( S 2 n + 1 ) = S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F k ( ξ ) G k ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F k L p ( S 2 n + 1 ) G k L q ( S 2 n + 1 ) .

Then, by equation (3.17) and dominated convergence theorem, we obtain

limsup p p α , q q β C Q , α , β , λ , p * S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F k ( ξ ) G k ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F k L p ( S 2 n + 1 ) G k L q ( S 2 n + 1 ) .

Letting k , we arrive at

(3.18) limsup p p α , q q β C Q , α , β , λ , p * N α , β , λ .

Therefore, by equations (3.16) and (3.18), we have

limsup p p α , q q β C Q , α , β , λ , p * = N α , β , λ .

Then, using Hölder’s inequality yields

N α , β , λ S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F p ( ξ ) G q ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η F p L p α ( S 2 n + 1 ) G q L q β ( S 2 n + 1 ) S 2 n + 1 S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 F p ( ξ ) G q ( η ) ( 1 η n + 1 2 ) β 2 d ξ d η S 2 n + 1 1 p α 1 p + 1 q β 1 q N α , β , λ , as p p α , q q β .

Hence, we obtain equation (3.15).□

Now we are ready to give the proof of Theorem 3.1.

Proof of Theorem 3.1

We start our proof by letting { F p , G q } C γ ( S 2 n + 1 ) × C γ ( S 2 n + 1 ) be a minimizing sequence of N α , β , λ . Then, { F p , G q } satisfies equation (3.5). By the scale and rotation invariance, without loss of generality, we assume that F p ( W ) = max ξ S 2 n + 1 F p ( ξ ) with W = ( 1 , 0 , , 0 ) .

There are two possible cases as follows.

Case 1. For some subsequences p j p α and q j q β , max { max ξ S 2 n + 1 F p j , max ξ S 2 n + 1 G q j } is uniformly bounded.

By repeating the same argument as equations (3.6) and (3.14), we deduce that { F p j } and { G q j } are uniformly bounded and equicontinuous on S 2 n + 1 . By equation (3.5), we know that there exists some constant C > 0 (independent of p j and q j ) such that F p j , G q j C . Then, it follows from Arelà-Ascoli theorem that there exist two subsequences of { F p j } and { G q j } (still denoted by { F p j } and { G q j } ) and two nonnegative functions F , G C γ ( S 2 n + 1 ) such that

F p j F , G q j G uniformly on S 2 n + 1 .

This implies

S 2 n + 1 F p α ( ξ ) d ξ = lim p j p α S 2 n + 1 F p j p j ( ξ ) d ξ = 1 ,

S 2 n + 1 G q β ( ξ ) d ξ = lim q j q β S 2 n + 1 G q j q j ( ξ ) d ξ = 1 .

Then, combining equation (3.5) with Lemma 3.4 gives

N α , β , λ F p α 1 ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 G ( η ) d η , N α , β , λ G q β 1 ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) β 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) α 2 F ( η ) d η

as j + . It follows that F and G are minimizers.

Case 2. For any subsequences p j p α and q j q β , F p j ( W ) + or max ξ S 2 n + 1 G q j + . Without loss of generality, we assume that F p j ( W ) + .

Case (i). limsup j + F p j ( W ) max ξ S 2 n + 1 G q j = + . Then, there exist two subsequences of p j and q j (still denoted by p j and q j ) such that F p j ( W ) + and F p j ( W ) max ξ S 2 n + 1 G q j + . Let ϕ j = F p j p j 1 , ψ j = G q j q j 1 , and noting that F p j L p j ( S 2 n + 1 ) = G q j L q j ( S 2 n + 1 ) = 1 , we derive

(3.19) S 2 n + 1 ϕ j p j ( ξ ) d ξ = S 2 n + 1 ψ j q j ( ξ ) d ξ = 1 .

From this and equation (3.5), we have

C Q , α , β , λ , p j * ϕ j ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) α 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) β 2 ψ j q j 1 ( η ) d η , C Q , α , β , λ , p j * ψ j ( ξ ) = S 2 n + 1 ( 1 ξ n + 1 2 ) β 2 1 ξ η ¯ λ 2 ( 1 η n + 1 2 ) α 2 ϕ j p j 1 ( ξ ) d η .

Then, applying Cayley transform and dilation on H n gives

(3.20) C Q , α , β , λ , p j * ( ( 1 + z 2 ) 2 + ( 2 t ) 2 ) λ + 2 α 4 ϕ j ( C ( δ ( u ) ) ) = 2 λ 2 + Q 1 2 Q + λ + 2 α + 2 β 2 H n z α v 1 u λ z β ψ j q j 1 ( C ( δ ( u ) ) ) ( ( 1 + z 2 ) 2 + ( 2 t ) 2 ) λ + 2 β + 2 Q 4 d v , C Q , α , β , λ , p j * ( ( 1 + z 2 ) 2 + ( 2 t ) 2 ) λ + 2 β 4 ψ j ( C ( δ ( u ) ) ) = 2 λ 2 + Q 1 2 Q + λ + 2 α + 2 β 2 H n z β v 1 u λ z α ϕ j p j 1 ( C ( δ ( u ) ) ) ( ( 1 + z 2 ) 2 + ( 2 t ) 2 ) λ + 2 α + 2 Q 4 d v .

Now we take = j such that j ( 2 Q + λ + 2 α + 2 β ) q j 2 ( p j 1 ) ( q j 1 ) 2 ( 1 + j 2 ) λ + 2 α 2 ϕ j ( C ( 1 , 0 , 0 ) ) = 1 and let

(3.21) Φ j ( u ) = j ( 2 Q + λ + 2 α + 2 β ) q j 2 ( p j 1 ) ( q j 1 ) 2 ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) λ + 2 α 4 ϕ j ( C ( δ j ( u ) ) ) , Ψ j ( u ) = j ( 2 Q + λ + 2 α + 2 β ) p j 2 ( p j 1 ) ( q j 1 ) 2 ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) λ + 2 β 4 ψ j ( C ( δ j ( u ) ) ) .

By a direct computation, one finds that Φ j and Ψ j satisfy the following renormalized equations:

(3.22) C Q , α , β , λ , p j * Φ j ( u ) = 2 λ 2 + Q 1 H n z α v 1 u λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v , C Q , α , β , λ , p j * Ψ j ( u ) = 2 λ 2 + Q 1 H n z β v 1 u λ z α Φ j p j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 α ) p j + 2 Q 4 d v .

Note that C ( 1 , 0 , 0 ) = ( 1 , 0 , 0 ) . We also have Φ j ( u ) Φ j ( 1 , 0 , , 0 ) = 1 , and for p j q j ,

(3.23) Ψ j ( u ) j ( 2 Q + λ + 2 α + 2 β ) p j 2 ( p j 1 ) ( q j 1 ) 2 min ξ S 2 n + 1 ψ j = j ( 2 Q + λ + 2 α + 2 β ) p j 2 ( p j 1 ) ( q j 1 ) 2 ( 1 + j 2 ) λ + 2 α 2 min ξ S 2 n + 1 ψ j ϕ j ( C ( 1 , 0 , 0 ) ) + uniformly for any u H n as j + .

We claim that there exists a constant C 1 1 such that for any u H n ,

(3.24) 1 C 1 ( 1 + u λ ) Φ j ( u ) z α C 1 ( 1 + u λ ) .

Once the inequality (3.24) holds and by equations (1.3) and (3.22) and polar decomposition, we find

(3.25) C Q , α , β , λ , p j * Ψ j ( 1 , 0 , , 0 ) C H n ( 1 + v λ ) z α Φ j p j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 α ) p j + 2 Q 4 d v C H n \ B ( 0 , 1 ) z α p j v λ p j d v + C B ( 0 , 1 ) ( 1 + v λ ) z α Φ j p j 1 ( v ) d v C ω n 1 ρ ( α + λ ) p j + Q 1 d ρ 0 π ( sin θ ) α p j 2 + n 1 d θ + C B ( 0 , 1 ) z α ( 1 + v λ ) d v = C ω n Q + ( α + λ ) p j B ( 1 2 , 2 n + α p j 4 ) + C ,

where ω n is defined as in Theorem 1.1. Here, we use the fact ( α + λ ) p j + Q < 0 and Φ j ( u ) 1 in the third inequality. This contradicts equation (3.23), and hence Case (i) is impossible.

In the last part of Case (i), we prove that the inequality (3.24) holds. Using equations (3.19) and (3.21) gives

S 2 n + 1 ϕ j p j ( ξ ) d ξ = H n j ( 2 Q + λ + 2 α + 2 β ) p j q j 2 ( p j 1 ) ( q j 1 ) 2 ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 α ) p j 4 Φ j p j ( u ) d u = 1 ,

S 2 n + 1 ψ j q j ( ξ ) d ξ = H n j ( 2 Q + λ + 2 α + 2 β ) p j q j 2 ( p j 1 ) ( q j 1 ) 2 ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j 4 Ψ j q j ( u ) d u = 1 .

This implies

meas { u H n Φ j ( u ) < + } > 0 , meas { v H n Ψ j ( v ) < + } > 0 .

Then, there exist R > 1 large enough and a measurable set E such that

E { v H n Ψ j ( v ) < R } B ( 0 , R )

with meas ( E ) > 1 R .

On the one hand, for any u H n , we have

C Q , α , β , λ , p j * Φ j ( u ) = 2 λ 2 + Q 1 H n z α v 1 u λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v 2 λ 2 + Q 1 E z α v 1 u λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v C R q j 1 z α E v 1 u λ z β d v .

It follows that there exists a constant C 1 1 such that

Φ j ( u ) z α 1 C 1 ( 1 + u λ ) u H n .

Similarly, there exists a constant C 2 1 such that

(3.26) Ψ j ( u ) z β 1 C 2 ( 1 + u λ ) u H n .

On the other hand, we know that

(3.27) C Q , α , β , λ , p j * = C Q , α , β , λ , p j * Φ j ( 1 , 0 , , 0 ) = 2 λ 2 + Q 1 H n v 1 ( 1 , 0 , , 0 ) λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v < + .

To simplify, we write u ˜ = ( 1 , 0 , , 0 ) . Then, combining equations (3.26) and (3.27) gives

(3.28) H n ( 1 + v λ ) z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v C { 0 < v < 1 2 } v 1 u ˜ λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v + C { v > 2 } v 1 u ˜ λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v + { 1 2 < v < 2 } ( 1 + v λ ) z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v < + .

It follows that for any u H n ,

(3.29) Φ j ( u ) z α ( 1 + u λ ) = 2 λ 2 + Q 1 H n v 1 u λ z β Ψ j q j 1 ( v ) ( 1 + u λ ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v C H n ( 1 + v λ ) z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v < + .

Combining equations (3.26) and (3.29), we obtain equation (3.24).

Case (ii) limsup j + F p j ( W ) max ξ S 2 n + 1 G q j = 0 . Then, there exist two subsequences of p j and q j (still denoted by p j and q j ) such that F p j ( W ) + and F p j ( W ) max ξ S 2 n + 1 G q j 0 . This implies that max ξ S 2 n + 1 G q j + . Similar to Case (i), we know that Case (ii) does not hold.

Case (iii). limsup j + F p j ( W ) max ξ S 2 n + 1 G q j = c 0 ( 0 , + ) . Then, there exist two subsequences of p j and q j (still denoted by p j and q j ) such that F p j ( W ) + , max ξ S 2 n + 1 G q j + and F p j ( W ) max ξ S 2 n + 1 G q j c 0 . By repeating the similar argument as in Case (i), we choose { F j , G j } defined as in equation (3.21), and hence F j and G j satisfy equation (3.22). Moreover, Φ j ( u ) Φ j ( 1 , 0 , , 0 ) = 1 and for p j q j ,

Ψ j ( u ) j ( 2 Q + λ + 2 α + 2 β ) p j 2 ( p j 1 ) ( q j 1 ) 2 min ξ S 2 n + 1 ψ j = j ( 2 Q + λ + 2 α + 2 β ) p j 2 ( p j 1 ) ( q j 1 ) 2 ( 1 + j 2 ) λ + 2 α 2 min ξ S 2 n + 1 ψ j ϕ j ( C ( 1 , 0 , 0 ) ) c 1 > 0 uniformly for any u H n as j + .

Therefore, Ψ j ( u ) has uniformly lower bound c 1 > 0 .

Repeating the arguments as equation (3.24), we also deduce that there exist constants C 3 , C 4 1 such that for any u H n ,

(3.30) 1 C 3 ( 1 + u λ ) Φ j ( u ) z α C 3 ( 1 + u λ ) ,

(3.31) 1 C 4 ( 1 + u λ ) Ψ j ( u ) z β C 4 ( 1 + u λ ) .

Then, for any given R 0 > 0 and u B ( 0 , R 0 ) , using equations (3.28) and (3.31) and the lower bound of Ψ j ( u ) gives

C Q , α , β , λ , p j * Φ j ( u ) = 2 λ 2 + Q 1 H n z α v 1 u λ z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v C z α H n \ B ( 0 , u ) ( 1 + v λ ) z β Ψ j q j 1 ( v ) ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 β ) q j + 2 Q 4 d v + C z α B ( 0 , u ) ( 1 + v λ ) z β d v C z α .

Thus, Φ j ( u ) is uniformly bounded in B ( 0 , R 0 ) . Similarly, Ψ j ( u ) is uniformly bounded in B ( 0 , R 0 ) .

A similar argument as equation (3.14) also shows that Φ j ( u ) and Ψ j ( u ) are equicontinuous in B ( 0 , R 0 ) ¯ . By Arzelà-Ascoli theorem, we know that there exist two subsequences of { Φ j } and { Ψ j } (still denoted by { Φ j } and { Ψ j } ) and two functions Φ and Ψ with lower bound C > 0 such that

Φ j Φ and Ψ j Ψ as j + uniformly in B ( 0 , R 0 ) .

Since R 0 is arbitrary, we can conclude that Φ ( u ) and Ψ ( u ) satisfy

N α , β , λ Φ ( u ) = 2 λ 2 + Q 1 H n z α v 1 u λ z β Ψ q β 1 ( v ) d v N α , β , λ Ψ ( u ) = 2 λ 2 + Q 1 H n z β v 1 u λ z α Φ p α 1 ( v ) d v .

Since

1 = S 2 n + 1 ϕ j p j ( ξ ) d ξ = H n j Q ( 2 Q + λ + 2 α + 2 β ) p j q j 2 ( p j 1 ) ( q j 1 ) 2 ( ( 1 + j z 2 ) 2 + ( j 2 t ) 2 ) ( λ + 2 α ) p j 4 Φ j p j ( u ) d u H n Φ j p j ( u ) d u

and

Φ j p j Φ p α

uniformly on any compact domain, we know from equation (3.30) that

H n Φ p α ( u ) d u = lim j + H n Φ j p j ( u ) d u 1 .

Similarly, we have

H n Ψ q β ( u ) d u = lim j + H n Ψ j q j ( u ) d u 1 .

Now we set f ( u ) = Φ p α 1 ( u ) and g ( v ) = Ψ q β 1 ( u ) , then

H n f p α ( u ) d u 1 , H n g q β ( u ) d u 1 ,

and F and G satisfy

N α , β , λ f p α 1 ( u ) = 2 λ 2 + Q 1 H n z α v 1 u λ z β Ψ q β 1 ( v ) d v , N α , β , λ g q β 1 ( u ) = 2 λ 2 + Q 1 H n z β v 1 u λ z α Φ p α 1 ( v ) d v .

Since p α , q β < 1 , using the Cayley transform, we have

N α , β , λ 2 = 2 λ 2 + Q 1 H n H n z α v 1 u λ z β f ( u ) g ( v ) d u d v 2 H n f p α ( u ) d u H n g q β ( v ) d v 2 λ 2 + Q 1 H n H n z α v 1 u λ z β f ( u ) g ( v ) d u d v 2 H n f p α ( u ) d u 2 p α H n g q β ( v ) d v 2 q β = S 2 n + 1 S 2 n + 1 ξ α 1 ξ η ¯ λ 2 F ( ξ ) G ( η ) η β d ξ d η 2 H n F p α ( u ) d u 2 p α H n G q β ( v ) d v 2 q β .

Therefore, { F ( ξ ) , G ( η ) } is a pair of minimizers of sharp constant N α , β , λ . This completes the proof of Theorem 3.1.□

Proof of Theorem 1.4

Theorem 1.4 is an easy consequence of Theorem 3.1.□

Acknowledgments

The authors would like to thank the referee for his/her careful reading of the manuscript and valuable suggestions.

  1. Funding information: This project is supported by the National Natural Science Foundation of China (Grant No. 12101380), China Postdoctoral Science Foundation (Grant No. 2021M700086), Youth Innovation Team of Shaanxi Universities, and the Fundamental Research Funds for the Central Universities (Grant Nos. GK202307001 and GK202202007).

  2. Conflict of interest: The authors state that there is no conflict of interest.

References

[1] W. Beckner, Sharp inequalities and geometric manifolds, in: J. Fourier Anal. Appl. 3 (1997), 825–836, Special Issue: Proceedings of the Conference Dedicated to Professor Miguel de Guzmán, El Escorial, 1996. 10.1007/BF02656488Search in Google Scholar

[2] W. Beckner, Weighted inequalities and Stein-Weiss potentials, Forum Math. 20 (2008), 587–606. 10.1515/FORUM.2008.030Search in Google Scholar

[3] W. Beckner, Pitt’s inequality with sharp convolution estimates, Proc. Amer. Math. Soc. 136 (2008), 1871–1885. 10.1090/S0002-9939-07-09216-7Search in Google Scholar

[4] E. Carlen and M. Loss, Extremals of functionals with competing symmetries, J. Funct. Anal. 88 (1990), 437–456. 10.1016/0022-1236(90)90114-ZSearch in Google Scholar

[5] L. Chen, G. Lu, and C. Tao, Reverse Stein-Weiss inequalities on the upper half space and the existence of their extremals, Adv. Nonlinear Stud. 19 (2019), 475–494. 10.1515/ans-2018-2038Search in Google Scholar

[6] L. Chen, Z. Liu, G. Lu, and C. Tao, Reverse Stein-Weiss inequalities and existence of their extremal functions, Trans. Amer. Math. Soc. 370 (2018), 8429–8450. 10.1090/tran/7273Search in Google Scholar

[7] L. Chen, G. Lu, and C. Tao, Existence of extremal functions for the Stein-Weiss inequalities on the Heisenberg group, J. Funct. Anal. 277 (2019), 1112–1138. 10.1016/j.jfa.2019.01.002Search in Google Scholar

[8] W. Dai, Y. Hu, and Z. Liu, Sharp reversed Hardy-Littlewood-Sobolev inequality with extended kernel, Studia Math. 271 (2023), 1–38, https://doi.org/10.4064/sm220323-26-1. Search in Google Scholar

[9] J. Dou, Weighted Hardy-Littlewood-Sobolev inequalities on the upper half space, Comm. Cont. Math. 18 (2016), 1550067. 10.1142/S0219199715500674Search in Google Scholar

[10] J. Dou, Q. Guo, and M. Zhu, Subcritical approach to sharp Hardy-Littlewood-Sobolev type inequalities on the upper half space, Adv. Math. 312 (2017), 1–45. 10.1016/j.aim.2017.03.007Search in Google Scholar

[11] J. Dou, Q. Guo, and M. Zhu, Negative power nonlinear integral equations on bounded domains, J. Differential Equations 269 (2020), 10527–10557. 10.1016/j.jde.2020.07.021Search in Google Scholar

[12] X. Han, Existence of maximizers for Hardy-Littlewood-Sobolev inequalities on the Heisenberg group, Indiana Univ. Math. J. 62 (2013), 737–751. 10.1512/iumj.2013.62.4976Search in Google Scholar

[13] J. Dou, Y. Hu, and J. Ma, Reversed Hardy-Littlewood-Sobolev inequalities with vertical weights on the upper half space, preprint, arXiv:2311.03933.Search in Google Scholar

[14] J. Dou and M. Zhu, Sharp Hardy-Littlewood-Sobolev inequality on the upper half space, Int. Math. Res. Not. 3 (2015), 651–687. 10.1093/imrn/rnt213Search in Google Scholar

[15] J. Dou and M. Zhu, Reversed Hardy-Littewood-Sobolev inequality, Int. Math. Res. Not. 19 (2015), 9696–9726. 10.1093/imrn/rnu241Search in Google Scholar

[16] J. Dou and M. Zhu, Integral equations on bounded domains, J. Funct. Anal. 277 (2019), 111–134. 10.1016/j.jfa.2018.05.020Search in Google Scholar

[17] G. B. Folland and E. M. Stein, Estimates for the ∂¯b complex and analysis on the Heisenberg group, Comm. Pure Appl. Math. 27 (1974), 429–522. 10.1002/cpa.3160270403Search in Google Scholar

[18] R. L. Frank and E. H. Lieb, Inversion positivity and the sharp Hardy-Littlewood-Sobolev inequality, Calc. Var. Partial Differential Equations 39 (2010), 85–99. 10.1007/s00526-009-0302-xSearch in Google Scholar

[19] R. L. Frank and E. H. Lieb, A new rearrangement-free proof of the sharp Hardy-Littlewood-Sobolev inequality, spectral theory, function spaces and inequalities, in: B. M. E. A Brown, ed., Oper. Theory Adv. Appl. vol. 219, Birkh auser, Basel, 2012, pp. 55–67. 10.1007/978-3-0348-0263-5_4Search in Google Scholar

[20] R. L. Frank and E. H. Lieb, Sharp constants in several inequalities on the Heisenberg group, Ann. Math. 176 (2012), 349–381. 10.4007/annals.2012.176.1.6Search in Google Scholar

[21] M. Gluck, Subcritical approach to conformally invariant extension operators on the upper half space, J. Funct. Anal. 278 (2020), no. 1, 108082, 46 pp. 10.1016/j.jfa.2018.08.012Search in Google Scholar

[22] X. Han and G. Lu, Stein-Weiss inequalities and integral systems on the Heisenberg group, Nonlinear Anal. 75 (2012), 4296–4314. 10.1016/j.na.2012.03.017Search in Google Scholar

[23] Y. Han and S. Zhang, Reversed Hardy-Littlewood-Sobolev inequality on Heisenberg group Hn and CR sphere S2n+1, arXiv:2202.09241. Search in Google Scholar

[24] Y. Han and M. Zhu, Hardy-Littlewood-Sobolev inequalities on compact Riemannian manifolds and applications, J. Differential Equations 260 (2016), 1–25. 10.1016/j.jde.2015.06.032Search in Google Scholar

[25] F. Hang and X. Wang, A simpler proof of Frank and Lieb’s sharp inequality on the Heisenberg group, arXiv: 2211.10301v1. Search in Google Scholar

[26] G. H. Hardy and J. E. Littlewood, Some properties of fractional integrals, Math. Z. 27 (1928), 565–606. 10.1007/BF01171116Search in Google Scholar

[27] I. W. Herbst, Spectral theory of the operator (p2+m2)1⁄2−Ze2⁄r2, Comm. Math. Phys. 53 (1977), no. 3, 285–294. 10.1007/BF01609852Search in Google Scholar

[28] A. Kassymov, M. Ruzhansky, and D. Suragan, Reverse Stein-Weiss, Hardy-Littlewood-Sobolev, Hardy, Sobolev and Caffarelli-Kohn-Nirenberg inequalities on homogeneous groups, Forum Math. 34 (2022), 1147–1158. 10.1515/forum-2021-0110Search in Google Scholar

[29] A. Kassymov, M. Ruzhansky, and D. Suragan, Reverse integral Hardy inequality on metric measure spaces, Ann. Fenn. Math. 47 (2022), 39–55. 10.54330/afm.112455Search in Google Scholar

[30] E. H. Lieb, Sharp constants in the Hardy-Littlewood-Sobolev and related inequalities, Ann. Math. 118 (1983), 349–374. 10.2307/2007032Search in Google Scholar

[31] Q. A. Ngô and V. H. Nguyen, Sharp reversed Hardy-Littlewood-Sobolev inequality on Rn, Israel J. Math. 220 (2017), 189–223. 10.1007/s11856-017-1515-xSearch in Google Scholar

[32] Q. A. Ngô and V. H. Nguyen, Sharp reversed Hardy-Littlewood-Sobolev inequality on the half space R+n, Int. Math. Res. Not. 2017 (2017), no. 20, 6187–6230. 10.1093/imrn/rnw108Search in Google Scholar

[33] S. L. Sobolev, On a theorem in functional analysis (in Russian), Mat. Sb 4 (1938), 471–497. Search in Google Scholar

[34] E. M. Stein and G. Weiss, Fractional integrals on n-dimensional Euclidean space, J. Math. Mech. 7 (1958), 503–514. 10.1512/iumj.1958.7.57030Search in Google Scholar

[35] C. Tao, Reversed Stein-Weiss inequalities with Poisson-type kernel and qualitative analysis of extremal functions, Adv. Nonlinear Stud. 21 (2021), 167–187. 10.1515/ans-2020-2112Search in Google Scholar

[36] S. Zhang and Y. Han, Rearrangement free method for Hardy-Littlewood-Sobolev inequalities on Sn, Anal. Theory Appl. 3x (2022), 1–26. Search in Google Scholar

Received: 2023-06-21
Revised: 2023-10-15
Accepted: 2023-10-26
Published Online: 2023-11-20

© 2023 the author(s), published by De Gruyter

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

Articles in the same Issue

  1. Regular Articles
  2. On sufficient “local” conditions for existence results to generalized p(.)-Laplace equations involving critical growth
  3. On the critical Choquard-Kirchhoff problem on the Heisenberg group
  4. On the local behavior of local weak solutions to some singular anisotropic elliptic equations
  5. Viscosity method for random homogenization of fully nonlinear elliptic equations with highly oscillating obstacles
  6. Double-phase parabolic equations with variable growth and nonlinear sources
  7. Logistic damping effect in chemotaxis models with density-suppressed motility
  8. Bifurcation diagrams of one-dimensional Kirchhoff-type equations
  9. Standing wave solution for the generalized Jackiw-Pi model
  10. Weak-strong uniqueness and energy-variational solutions for a class of viscoelastoplastic fluid models
  11. Eigenvalue inequalities for the buckling problem of the drifting Laplacian of arbitrary order
  12. Homoclinic solutions for a differential inclusion system involving the p(t)-Laplacian
  13. Equivalence between a time-fractional and an integer-order gradient flow: The memory effect reflected in the energy
  14. Bautin bifurcation with additive noise
  15. Small solitons and multisolitons in the generalized Davey-Stewartson system
  16. Nonstationary Poiseuille flow of a non-Newtonian fluid with the shear rate-dependent viscosity
  17. A global compactness result with applications to a Hardy-Sobolev critical elliptic system involving coupled perturbation terms
  18. On a strongly damped semilinear wave equation with time-varying source and singular dissipation
  19. Singularity for a nonlinear degenerate hyperbolic-parabolic coupled system arising from nematic liquid crystals
  20. Stability of stationary solutions to the three-dimensional Navier-Stokes equations with surface tension
  21. Uniform decay estimates for the semi-linear wave equation with locally distributed mixed-type damping via arbitrary local viscoelastic versus frictional dissipative effects
  22. Symmetry and nonsymmetry of minimal action sign-changing solutions for the Choquard system
  23. Periodic and quasi-periodic solutions of a four-dimensional singular differential system describing the motion of vortices
  24. Multiple nontrivial solutions of superlinear fractional Laplace equations without (AR) condition
  25. Existence and blow-up of solutions in Hénon-type heat equation with exponential nonlinearity
  26. Existence and concentration behavior of positive solutions to Schrödinger-Poisson-Slater equations
  27. On the fractional Korn inequality in bounded domains: Counterexamples to the case ps < 1
  28. Gradient estimates for nonlinear elliptic equations involving the Witten Laplacian on smooth metric measure spaces and implications
  29. Dynamical analysis of a reaction–diffusion mosquito-borne model in a spatially heterogeneous environment
  30. Existence and multiplicity of solutions for a quasilinear system with locally superlinear condition
  31. Nontrivial solution for Klein-Gordon equation coupled with Born-Infeld theory with critical growth
  32. Modeling Wolbachia infection frequency in mosquito populations via a continuous periodic switching model
  33. Infinitely many localized semiclassical states for nonlinear Kirchhoff-type equation
  34. Fujita-type theorems for a quasilinear parabolic differential inequality with weighted nonlocal source term
  35. Approximations of center manifolds for delay stochastic differential equations with additive noise
  36. Periodic solutions to a class of distributed delay differential equations via variational methods
  37. Groundstate for the Schrödinger-Poisson-Slater equation involving the Coulomb-Sobolev critical exponent
  38. Existence of nontrivial solutions for the Klein-Gordon-Maxwell system with Berestycki-Lions conditions
  39. Global Sobolev regular solution for Boussinesq system
  40. Normalized solutions for the p-Laplacian equation with a trapping potential
  41. Nonlinear elliptic–parabolic problem involving p-Dirichlet-to-Neumann operator with critical exponent
  42. Blow-up for compressible Euler system with space-dependent damping in 1-D
  43. High energy solutions of general Kirchhoff type equations without the Ambrosetti-Rabinowitz type condition
  44. On the dynamics of grounded shallow ice sheets: Modeling and analysis
  45. A survey on some vanishing viscosity limit results
  46. Blow-up for logarithmic viscoelastic equations with delay and acoustic boundary conditions
  47. Generalized Liouville theorem for viscosity solutions to a singular Monge-Ampère equation
  48. Front propagation in a double degenerate equation with delay
  49. Positive solutions for a class of singular (pq)-equations
  50. Higher integrability for anisotropic parabolic systems of p-Laplace type
  51. The Poincaré map of degenerate monodromic singularities with Puiseux inverse integrating factor
  52. On a system of multi-component Ginzburg-Landau vortices
  53. Existence and concentration of solutions to Kirchhoff-type equations in ℝ2 with steep potential well vanishing at infinity and exponential critical nonlinearities
  54. Multiplicity results for fractional Schrödinger-Kirchhoff systems involving critical nonlinearities
  55. On double phase Kirchhoff problems with singular nonlinearity
  56. Estimates for eigenvalues of the Neumann and Steklov problems
  57. Nodal solutions with a prescribed number of nodes for the Kirchhoff-type problem with an asymptotically cubic term
  58. Dirichlet problems involving the Hardy-Leray operators with multiple polars
  59. Incompressible limit for compressible viscoelastic flows with large velocity
  60. Characterizations for the genuine Calderón-Zygmund operators and commutators on generalized Orlicz-Morrey spaces
  61. Non-local gradients in bounded domains motivated by continuum mechanics: Fundamental theorem of calculus and embeddings
  62. Noncoercive parabolic obstacle problems
  63. Touchdown solutions in general MEMS models
  64. Existence and nonexistence of nontrivial solutions for the Schrödinger-Poisson system with zero mass potential
  65. Short-time existence of a quasi-stationary fluid–structure interaction problem for plaque growth
  66. Fine bounds for best constants of fractional subcritical Sobolev embeddings and applications to nonlocal PDEs
  67. Symmetries of Ricci flows
  68. Asymptotic behavior of solutions of a second-order nonlinear discrete equation of Emden-Fowler type
  69. On the topological gradient method for an inverse problem resolution
  70. Supersolutions to nonautonomous Choquard equations in general domains
  71. Uniform complex time heat Kernel estimates without Gaussian bounds
  72. Global existence for time-dependent damped wave equations with nonlinear memory
  73. Normalized solutions for a critical fractional Choquard equation with a nonlocal perturbation
  74. Reversed Hardy-Littlewood-Sobolev inequalities with weights on the Heisenberg group
  75. Lamé system with weak damping and nonlinear time-varying delay
  76. Global well posedness for the semilinear edge-degenerate parabolic equations on singular manifolds
  77. Blowup in L1(Ω)-norm and global existence for time-fractional diffusion equations with polynomial semilinear terms
  78. Boundary regularity results for minimisers of convex functionals with (p, q)-growth
  79. Parametric singular double phase Dirichlet problems
  80. Special Issue on Nonlinear analysis: Perspectives and synergies
  81. Editorial to Special issue “Nonlinear analysis: Perspectives and synergies”
  82. Identification of discontinuous parameters in double phase obstacle problems
  83. Global boundedness to a 3D chemotaxis-Stokes system with porous medium cell diffusion and general sensitivity
  84. On regular solutions to compressible radiation hydrodynamic equations with far field vacuum
  85. On Cauchy problem for fractional parabolic-elliptic Keller-Segel model
  86. The evolution of immersed locally convex plane curves driven by anisotropic curvature flow
  87. Stability of combination of rarefaction waves with viscous contact wave for compressible Navier-Stokes equations with temperature-dependent transport coefficients and large data
  88. On concave perturbations of a periodic elliptic problem in R2 involving critical exponential growth
  89. Existence of solutions for a double-phase variable exponent equation without the Ambrosetti-Rabinowitz condition
Downloaded on 10.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2023-0116/html
Scroll to top button