• 検索結果がありません。

1.Introduction ShanliYeandZhengyuanZhuo WeightedCompositionOperatorsfromHardytoZygmundTypeSpaces ResearchArticle

N/A
N/A
Protected

Academic year: 2022

シェア "1.Introduction ShanliYeandZhengyuanZhuo WeightedCompositionOperatorsfromHardytoZygmundTypeSpaces ResearchArticle"

Copied!
11
0
0

読み込み中.... (全文を見る)

全文

(1)

Volume 2013, Article ID 365286,10pages http://dx.doi.org/10.1155/2013/365286

Research Article

Weighted Composition Operators from Hardy to Zygmund Type Spaces

Shanli Ye and Zhengyuan Zhuo

Department of Mathematics, Fujian Normal University, Fuzhou 350007, China

Correspondence should be addressed to Shanli Ye; ye [email protected] Received 20 January 2013; Accepted 24 March 2013

Academic Editor: Yansheng Liu

Copyright © 2013 S. Ye and Z. Zhuo. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper aims at studying the boundedness and compactness of weighted composition operator between spaces of analytic functions. We characterize boundedness and compactness of the weighted composition operator𝑢𝐶𝜙from the Hardy spaces𝐻𝑝 to the Zygmund type spacesZ𝛼= {𝑓 ∈ 𝐻(𝐷) :sup𝑧∈𝐷(1 − |𝑧|2)𝛼|𝑓󸀠󸀠(𝑧)| < ∞}and the little Zygmund type spacesZ𝛼,0in terms of function theoretic properties of the symbols𝑢and𝜙.

1. Introduction

Let𝐷 = {𝑧 : |𝑧| < 1}be the open unit disk in the complex planeC and𝑇 = {𝑧 : |𝑧| = 1}its boundary, and𝐻(𝐷)denotes the set of all analytic functions on𝐷. An analytic self-map 𝜑 : 𝐷 → 𝐷induces the composition operator𝐶𝜑on𝐻(𝐷), defined by𝐶𝜑(𝑓) = 𝑓(𝜑(𝑧))for𝑓analytic on𝐷. It is a well- known consequence of Littlewood’s subordination principle that the composition operator𝐶𝜑is bounded on the classical Hardy𝐻𝑝 (0 < 𝑝 ≤ ∞)spaces, Bergman𝐴𝑝 (0 < 𝑝 ≤ ∞) spaces, and Bloch spaces (see, e.g., [1–4]).

Let 𝑢 be a fixed analytic function on the open unit disk. Define a linear operator𝑢𝐶𝜑 on the space of analytic functions on𝐷, called a weighted composition operator, by 𝑢𝐶𝜑𝑓 = 𝑢 ⋅ (𝑓 ∘ 𝜑), where𝑓is an analytic function on𝐷. We can regard this operator as a generalization of a multiplication operator and a composition operator. In recent years the weighted composition operator has received much attention and appears in various settings in the literature. For example, it is known that isometries of many analytic function spaces are weighted composition operators (see [5], for instance).

Their boundedness and compactness have been studied on various Banach spaces of analytic functions, such as Hardy, Bergman, BMOA, Bloch-type, and Zygmund spaces; see, for example, [6–11]. Also, it has been studied from one Banach space of analytic functions to another; one may see [12–23].

The purpose of this paper is to consider the weighted composition operators from the Hardy space𝐻𝑝 (0 < 𝑝 <

∞) to the Zygmund type spacesZ𝛼. Our main goal is to characterize boundedness and compactness of the operators 𝑢𝐶𝜑from𝐻𝑝toZ𝛼in terms of function theoretic properties of the symbols𝑢and𝜑.

Now we give a detailed definition of these spaces. For0 ≤ 𝑟 < 1,𝑓(𝑧) ∈ 𝐻(𝐷), we set

𝑀𝑝(𝑟, 𝑓) = ( 1 2𝜋∫2𝜋

0 󵄨󵄨󵄨󵄨󵄨𝑓 (𝑟𝑒𝑖𝜃)󵄨󵄨󵄨󵄨󵄨𝑝𝑑𝜃)

1/𝑝

, 0 < 𝑝 < ∞, 𝑀(𝑟, 𝑓) = max

0≤𝜃≤2𝜋󵄨󵄨󵄨󵄨󵄨𝑓 (𝑟𝑒𝑖𝜃)󵄨󵄨󵄨󵄨󵄨.

(1)

For0 < 𝑝 ≤ ∞, the Hardy space 𝐻𝑝 consists of those functions𝑓 ∈ 𝐻(𝐷), for which

󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝= sup

0≤𝑟<1𝑀𝑝(𝑟, 𝑓) < ∞. (2) It is well known that with norm (2) the𝐻𝑝space is a Banach space if1 ≤ 𝑝 ≤ ∞, for0 < 𝑝 < 1,𝐻𝑝space is a nonlocally convex topological vector space, and𝑑(𝑓, 𝑔) = ‖𝑓 − 𝑔‖𝑝𝑝is a complete metric for it. For more information about the𝐻𝑝 space, one may see these books, for example, [24,25].

(2)

For𝛼 > 0the𝛼-Bloch space𝛽𝛼 consists of all analytic functions𝑓defined on𝐷such that

󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝛽𝛼 =sup{(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑓󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 : 𝑧 ∈ 𝐷} , 0 < 𝛼 < +∞.

(3) The spaceZ𝛼consists of all analytic functions𝑓defined on 𝐷such that

󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩Z𝛼 =sup{(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑓󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 : 𝑧 ∈ 𝐷} , 0 < 𝛼 < +∞.

(4) When𝛼 = 1, it is called the Zygmund space. From a theorem by Zygmund (see [26, vol. I, p. 263] or [24, Theorem 5.3]), we see that𝑓 ∈Z1if and only if𝑓is continuous in the close unit disk𝐷 = {𝑧 : |𝑧| ≤ 1}and the boundary function𝑓(𝑒𝑖𝜃)such that

󵄨󵄨󵄨󵄨󵄨𝑓 (𝑒𝑖(𝜃+ℎ)) + 𝑓 (𝑒𝑖(𝜃−ℎ)) − 2𝑓 (𝑒𝑖𝜃)󵄨󵄨󵄨󵄨󵄨

ℎ < ∞. (5)

When𝛼 > 1, from Proposition 8 of [27], we know thatZ𝛼= 𝛽𝛼−1. Then the spaceZ𝛼 is called a Zygmund type space if 0 < 𝛼 ≤ 1. However, all results in this paper are valid for all Z𝛼spaces (𝛼 > 0). An analytic function𝑓 ∈ 𝐻(𝐷)is said to belong to the little Zymund type spaceZ𝛼,0which consists of all𝑓 ∈Z𝛼satisfying lim|𝑧| → 1(1 − |𝑧|2)𝛼|𝑓󸀠󸀠(𝑧)| = 0. It can be easily proved thatZ𝛼is a Banach space under the norm

󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩 = 󵄨󵄨󵄨󵄨𝑓 (0)󵄨󵄨󵄨󵄨 + 󵄨󵄨󵄨󵄨󵄨𝑓󸀠(0)󵄨󵄨󵄨󵄨󵄨 +󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩Z𝛼. (6) And the polynomials are norm-dense in closed subspace Z𝛼,0. For some other information on this space and some operators on it, see, for example, [28–31].

Throughout this paper, constants are denoted by𝐶,𝐶(𝑝), they are positive, and𝐶(𝑝)are only depending on𝑝and may differ from one occurrence to the another.

2. Auxiliary Results

In order to prove the main results of this paper. We need some auxiliary results. The first lemma is well known.

Lemma 1 (see [24, p. 65]). For𝑝 > 1, there exists a constant 𝐶(𝑝)such that

2𝜋

0

𝑑𝜃

|1 − 𝑧|𝑝 ≤ 𝐶 (𝑝)

(1 − |𝑧|2)𝑝−1 for every 𝑧 ∈ 𝐷. (7) Lemma 2. Suppose that0 < 𝑝 < ∞,𝑓 ∈ 𝐻𝑝; then

󵄨󵄨󵄨󵄨󵄨𝑓(𝑛)(𝑧)󵄨󵄨󵄨󵄨󵄨 ≤ 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

(1 − |𝑧|2)1/𝑝+𝑛 (8)

for every𝑧 ∈ 𝐷and all nonnegative integer𝑛 = 0, 1, 2, . . ..

Proof. We use induction on𝑛. The case𝑛 = 0holds because it is Exercise 5 in [25, p. 85]. Assume the case𝑛 = 𝑘holds. Fix

0 < 𝑟 < 1and let𝑔(𝑧) = 𝑓(𝑘)(𝑟𝑧). Then𝑔(𝑧)is in𝐻 ⊂ 𝛽1, and‖𝑔‖𝛽1 ≤ ‖𝑔‖. It follows that

(1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑔󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 = (1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑟𝑓(𝑘+1)(𝑟𝑧)󵄨󵄨󵄨󵄨󵄨

≤ 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

(1 − |𝑟𝑧|2)1/𝑝+𝑘

≤ 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

(1 − |𝑧|2)1/𝑝+𝑘.

(9)

Let𝑟 → 1; we have

󵄨󵄨󵄨󵄨󵄨𝑓(𝑘+1)(𝑧)󵄨󵄨󵄨󵄨󵄨 ≤ 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

(1 − |𝑧|2)1/𝑝+𝑘+1. (10) Then the case𝑛 = 𝑘 + 1holds. Hence (8) holds.

Lemma 3. For0 < 𝑝 < ∞, suppose𝑢𝐶𝜑 : 𝐻𝑝 → Z𝛼,0 is a bounded operator. Then𝑢𝐶𝜑 : 𝐻𝑝 → Z𝛼is a bounded operator.

This is obvious.

3. Boundedness of 𝑢𝐶

𝜑

from 𝐻

𝑝

(0 < 𝑝 < ∞) to Z

𝛼

and Z

𝛼,0

In this section we characterize bounded weighted composi- tion operators from the Hardy space𝐻𝑝 (0 < 𝑝 < ∞)to the Zygmund spacesZ𝛼.

Theorem 4. Let𝛼 > 0,0 < 𝑝 < ∞, and𝑢be an analytic function on the unit disc𝐷and𝜑an analytic self-map of𝐷.

Then 𝑢𝐶𝜑 is a bounded operator from𝐻𝑝 to the Zygmund spacesZ𝛼if and only if the following are satisfied:

sup𝑧∈𝐷

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 < ∞, (11) sup𝑧∈𝐷

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 < ∞, (12) sup𝑧∈𝐷

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 < ∞. (13) Proof . Suppose𝑢𝐶𝜑 is bounded from𝐻𝑝 to the Zygmund spacesZ𝛼. Then we can easily obtain the following results by taking𝑓(𝑧) = 1and𝑓(𝑧) = 𝑧in𝐻𝑝, respectively:

𝑢 ∈Z𝛼,

sup𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧) + 𝜑 (𝑧) 𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

< +∞.

(14)

(3)

By (14) and the boundedness of the function𝜑(𝑧), we get 𝐾1=sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 < +∞.

(15) Let𝑓(𝑧) = 𝑧2in𝐻𝑝again; in the same way we have

sup𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨󵄨 4𝜑(𝑧)𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑2(𝑧) 𝑢󸀠󸀠(𝑧) +2𝑢 (𝑧) (𝜑 (𝑧) 𝜑󸀠󸀠(𝑧) + (𝜑󸀠(𝑧))2)󵄨󵄨󵄨󵄨󵄨󵄨󵄨 < ∞.

(16) Using these facts and the boundedness of the function𝜑(𝑧) again, we get

𝐾2=sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨 < +∞. (17) Fix𝑎 ∈ 𝐷; we take the test functions

𝑓𝑎(𝑧) = − 1 − |𝑎|2

(1 − 𝑎𝑧)1+1/𝑝 + 2 (1 − |𝑎|2)2

(1 − 𝑎𝑧)2+1/𝑝 − (1 − |𝑎|2)3 (1 − 𝑎𝑧)3+1/𝑝

(18) for 𝑧 ∈ 𝐷. From Lemma 1we obtain that 𝑓𝑎 ∈ 𝐻𝑝 and sup𝑎‖𝑓𝑎𝑝 ≤ 𝐶(𝑝) < ∞ with a direct calculation. Since 𝑓𝑎(𝑎) = 0,𝑓𝑎󸀠(𝑎) = 0, and𝑓𝑎󸀠󸀠(𝑎) = −2𝑎2/(1 − |𝑎|2)2+1/𝑝, it follows that, for all𝜆 ∈ 𝐷with|𝜑(𝜆)| > 1/2, we have

𝐶󵄩󵄩󵄩󵄩𝑓𝑎󵄩󵄩󵄩󵄩 ≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑓𝑎󵄩󵄩󵄩󵄩󵄩≥sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑓𝑎)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

=sup

𝑧∈𝐷(1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓𝑎󸀠(𝜑 (𝑧)) + 𝑓𝑎󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) 𝑓𝑎(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨.

(19)

Let𝑎 = 𝜑(𝜆); it follows that 𝐶󵄩󵄩󵄩󵄩𝑓𝑎󵄩󵄩󵄩󵄩≥ (1 − |𝜆|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)) 𝑓𝜑(𝜆)󸀠 (𝜑 (𝜆)) + 𝑓𝜑(𝜆)󸀠󸀠 (𝜑 (𝜆)) (𝜑󸀠(𝜆))2𝑢 (𝜆)

+ 𝑢󸀠󸀠(𝜆) 𝑓𝜑(𝜆)(𝜑 (𝜆))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= (1−|𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨

(𝜑󸀠(𝜆))2𝑢 (𝜆) 2(𝜑 (𝜆))2 (1−󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)2+1/𝑝

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨

≥ 1 2

(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝜆) (𝜑󸀠(𝜆))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)2+1/𝑝 .

(20)

For all𝜆 ∈ 𝐷with|𝜑(𝜆)| ≤ 1/2, by (17), we have

sup𝜆∈𝐷

(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝜆) (𝜑󸀠(𝜆))2󵄨󵄨󵄨󵄨󵄨󵄨

1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨1/𝑝+2

≤ (4

3)1/𝑝+2sup

𝜆∈𝐷(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝜆) (𝜑󸀠(𝜆))2󵄨󵄨󵄨󵄨󵄨󵄨

< +∞.

(21)

Hence (12) holds.

Next, fix𝑎 ∈ 𝐷; we take another test functions

𝑔𝑎(𝑧)= −3 1 − |𝑎|2

(1−𝑎𝑧)1/𝑝+1+5 (1 − |𝑎|2)2

(1 − 𝑎𝑧)1/𝑝+2−2 (1−|𝑎|2)3 (1−𝑎𝑧)1/𝑝+3

(22)

for 𝑧 ∈ 𝐷. From Lemma1 we obtain that 𝑔𝑎 ∈ 𝐻𝑝 and sup𝑎‖𝑔𝑎𝑝 ≤ 𝐶(𝑝) < ∞with a direct calculation. Since 𝑔𝑎(𝑎) = 0,𝑔󸀠󸀠𝑎(𝑎) = 0, and𝑔󸀠𝑎(𝑎) = 𝑎/(1 − |𝑎|2)1/𝑝+1, it follows that, for all𝜆 ∈ 𝐷with|𝜑(𝜆)| > 1/2, we obtain that

𝐶󵄩󵄩󵄩󵄩𝑔𝑎󵄩󵄩󵄩󵄩≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑔𝑎(𝑧)󵄩󵄩󵄩󵄩󵄩

≥sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑔𝑎)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

=sup

𝑧∈𝐷(1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑔𝑎󸀠(𝜑 (𝑧)) + 𝑔󸀠󸀠𝑎 (𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) 𝑔𝑎(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

≥ (1 − |𝜆|2)𝛼

(4)

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)) 𝑔𝜑(𝜆)󸀠 (𝜑 (𝜆)) + 𝑔󸀠󸀠𝜑(𝜆)(𝜑 (𝜆)) (𝜑󸀠(𝜆))2𝑢 (𝜆)

+ 𝑢󸀠󸀠(𝜆) 𝑔𝜑(𝜆)(𝜑 (𝜆))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − |𝜆|2)𝛼

×󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨

(2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆))

× 𝜑 (𝜆)

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)1/𝑝+1

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨

≥ 1 2

(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)1/𝑝+1 .

(23)

For all𝜆 ∈ 𝐷with|𝜑(𝜆)| ≤ 1/2, by (15), we have

|𝜑(𝜆)sup|≤1/2

(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)1/𝑝+1

≤ (4

3)1/𝑝+1 sup

|𝜑(𝜆)|≤1/2(1 − |𝜆|2)𝛼

× 󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)󵄨󵄨󵄨󵄨󵄨

< ∞.

(24)

Hence (13) holds.

Finally, fix𝑎 ∈ 𝐷, and, for all𝑧 ∈ 𝐷, let

𝑎(𝑧) = −𝑝 + 3 𝑝 + 1

1 − |𝑎|2 (1 − 𝑎𝑧)1/𝑝+1 +2 (𝑝 + 3)

𝑝 + 2

(1 − |𝑎|2)2

(1 − 𝑎𝑧)1/𝑝+2 − (1 − |𝑎|2)3 (1 − 𝑎𝑧)1/𝑝+3.

(25)

From Lemma1 we obtain thatℎ𝑎 ∈ 𝐻𝑝 and sup𝑎‖ℎ𝑎𝑝 ≤ 𝐶(𝑝) < ∞ with a direct calculation. Since ℎ󸀠𝑎(𝑎) = 0,

󸀠󸀠𝑎(𝑎) = 0, andℎ𝑎(𝑎) = −2/(𝑝 + 1)(𝑝 + 2)(1 − |𝑎|2)1/𝑝, it follows that, for all𝜆 ∈ 𝐷, we obtain that

𝐶󵄩󵄩󵄩󵄩ℎ𝑎󵄩󵄩󵄩󵄩≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑎(𝑧)󵄩󵄩󵄩󵄩󵄩

≥sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑎)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

=sup

𝑧∈𝐷(1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) ℎ󸀠𝑎(𝜑 (𝑧)) + ℎ󸀠󸀠𝑎(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) ℎ𝑎(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

≥ (1 − |𝜆|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝜆) 𝑢󸀠(𝜆) + 𝜑󸀠󸀠(𝜆) 𝑢 (𝜆)) ℎ󸀠𝜑(𝜆)(𝜑 (𝜆)) + ℎ󸀠󸀠𝜑(𝜆)(𝜑 (𝜆)) (𝜑󸀠(𝜆))2𝑢 (𝜆)

+ 𝑢󸀠󸀠(𝜆) ℎ𝜑(𝜆)(𝜑 (𝜆))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝜆) ℎ𝜑(𝜆)(𝜑 (𝜆))󵄨󵄨󵄨󵄨󵄨

= 2

(𝑝 + 1) (𝑝 + 2)

(1 − |𝜆|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝜆)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝜆)󵄨󵄨󵄨󵄨2)1/𝑝 .

(26) Then (11) holds.

Conversely, suppose that (11), (12), and (13) hold. For𝑓 ∈ 𝐻𝑝, by Lemma2, we have the following inequality:

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓󸀠(𝜑 (𝑧)) + 𝑓󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) 𝑓 (𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

≤ (1 − |𝑧|2)𝛼

× 󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓󸀠(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨 + (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑓󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

+ (1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧) 𝑓 (𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨

(5)

≤ (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

≤ 𝐶󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝,

󵄨󵄨󵄨󵄨𝑢(0)𝑓(𝜑(0))󵄨󵄨󵄨󵄨 + 󵄨󵄨󵄨󵄨󵄨𝑢󸀠(0) 𝑓 (𝜑 (0))󵄨󵄨󵄨󵄨󵄨 +󵄨󵄨󵄨󵄨󵄨𝑢 (0) 𝑓󸀠(𝜑 (0))󵄨󵄨󵄨󵄨󵄨

≤ ( |𝑢 (0)|

(1 − 󵄨󵄨󵄨󵄨𝜑 (0)󵄨󵄨󵄨󵄨2)1/𝑝 + 󵄨󵄨󵄨󵄨󵄨𝑢󸀠(0)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (0)󵄨󵄨󵄨󵄨2)1/𝑝

+ |𝑢 (0)|

(1 − 󵄨󵄨󵄨󵄨𝜑 (0)󵄨󵄨󵄨󵄨2)1/𝑝+1) 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝.

(27)

This shows that𝑢𝐶𝜑is bounded. This completes the proof of Theorem4.

Theorem 5. Let𝛼 > 0,0 < 𝑝 < ∞, and𝑢be an analytic function on the unit disc𝐷and𝜑an analytic self-map of𝐷.

Then𝑢𝐶𝜑 : 𝐻𝑝 → Z𝛼,0is a bounded operator provided that the following are satisfied:

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 = 0, (28)

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0, (29)

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 = 0. (30) Conversely, if 𝑢𝐶𝜑 : 𝐻𝑝 → Z𝛼,0 is a bounded operator, then𝑢 ∈ Z𝛼,0,(11),(12), and(13)hold, and the following are satisfied:

|𝑧| → 1lim(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨 = 0, (31)

|𝑧| → 1lim(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 = 0. (32)

Proof. Assume that (28), (29), and (30) hold. Then for any𝜖 >

0, there is a constant𝛿,0 < 𝛿 < 1, such that𝛿 < |𝑧| < 1 implies

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 < 𝜖, (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 < 𝜖, (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 < 𝜖.

(33)

Then, for any𝑓 ∈ 𝐻𝑝, from Lemma2we obtain that (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓󸀠(𝜑 (𝑧)) + 𝑓󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) 𝑓 (𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

≤ (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

× 1

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+ (1−|𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨 1

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+(1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝 ≤ 3󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝𝜖.

(34) Hence𝑢𝐶𝜑𝑓 ∈ Z𝛼,0for all𝑓 ∈ Z𝛼,0. On the other hand, (25), (28), and (29) imply that (11), (12), and (13) hold; then 𝑢𝐶𝜑: 𝐻𝑝 → Z𝛼is bounded by Theorem4. So𝑢𝐶𝜑: 𝐻𝑝 → Z𝛼,0is bounded.

Conversely, assume that𝑢𝐶𝜑is bounded from𝐻𝑝to the little Zygmund type spaceZ𝛼,0. Then𝑢 = 𝑢𝐶𝜑1 ∈Z𝛼,0. Also 𝑢𝜑 = 𝑢𝐶𝜑𝑧 ∈Z𝛼,0; thus

(1−|𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧)+𝜑󸀠󸀠(𝑧) 𝑢 (𝑧) + 𝜑 (𝑧) 𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 󳨀→ 0 (|𝑧| 󳨀→ 1) .

(35) Since |𝜑| ≤ 1 and 𝑢 ∈ Z𝛼,0, we have lim|𝑧| → 1(1 −

|𝑧|2)𝛼|2𝜑󸀠(𝑧)𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧)𝑢(𝑧)| = 0. Hence (32) holds.

(6)

Similarly,𝑢𝐶𝜑𝑧2∈Z𝛼,0; then

(1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨4𝜑 (𝑧) 𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑2(𝑧) 𝑢󸀠󸀠(𝑧)

+2𝑢 (𝑧) (𝜑 (𝑧) 𝜑󸀠󸀠(𝑧) + (𝜑󸀠(𝑧))2)󵄨󵄨󵄨󵄨󵄨󵄨󵄨 󳨀→ 0 (|𝑧| 󳨀→ 1) .

(36)

By (32),|𝜑| ≤ 1, and𝑢 ∈ Z𝛼,0, we get that lim|𝑧| → 1(1 −

|𝑧|2)𝛼|𝑢(𝑧)(𝜑󸀠(𝑧))2| = 0; that is, (31) holds.

On the other hand, from Lemma3and Theorem4, we obtain that (11), (12), and (13) hold.

4. Compactness of 𝑢𝐶

𝜑

In order to prove the compactness of𝑢𝐶𝜑 from𝐻𝑝 to the Zygmund spacesZ𝛼, we require the following lemmas.

Lemma 6. Let 𝛼 > 0,0 < 𝑝 < ∞, and𝑢be an analytic function on the unit disc𝐷and𝜑an analytic self-map of𝐷.

Suppose that𝑢𝐶𝜑is a bounded operator from𝐻𝑝toZ𝛼. Then 𝑢𝐶𝜑is compact if and only if, for any bounded sequence{𝑓𝑛} in𝐻𝑝which converges to0uniformly on compact subsets of𝐷, one has‖𝑢𝐶𝜑(𝑓𝑛)‖ → 0as𝑛 → ∞.

The proof is similar to that of Proposition 3.11 in [32]. The details are omitted.

Theorem 7. Let𝛼 > 0, 0 < 𝑝 < ∞,𝑢be an analytic function on the unit disc𝐷and𝜑an analytic self-map of𝐷. Then𝑢𝐶𝜑 is a compact operator from𝐻𝑝toZ𝛼if and only if𝑢𝐶𝜑is a bounded operator and the following are satisfied:

(i) lim

|𝜑(𝑧)|→ 1

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 = 0, (ii) lim

|𝜑(𝑧)|→ 1

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0, (iii) lim

|𝜑(𝑧)| → 1

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 = 0.

(37)

Proof. Suppose that𝑢𝐶𝜑is compact from𝐻𝑝to the Zygmund type space Z𝛼. Let {𝑧𝑛} be a sequence in 𝐷 such that

|𝜑(𝑧𝑛)| → 1 as 𝑛 → ∞. If such a sequence does not exist, then (37) are automatically satisfied. Without loss of

generality we may suppose that|𝜑(𝑧𝑛)| > 1/2for all𝑛. We take the test functions

𝑓𝑛(𝑧) = − 1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2

(1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+1 + 2 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)2 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+2

− (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)3 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+3.

(38)

By a direct calculation, we may easily prove that {𝑓𝑛} converges to 0 uniformly on compact subsets of 𝐷 and sup𝑛‖𝑓𝑛𝑝 ≤ 𝐶(𝑝) < ∞. Then{𝑓𝑛}is a bounded sequence in𝐻𝑝which converges to0uniformly on compact subsets of 𝐷. Then lim𝑛 → ∞‖𝑢𝐶𝜑(𝑓𝑛)‖= 0by Lemma6. Note that

𝑓𝑛(𝜑 (𝑧𝑛)) = 0, 𝑓𝑛󸀠(𝜑 (𝑧𝑛)) = 0, 𝑓𝑛󸀠󸀠(𝜑 (𝑧𝑛)) = − 2𝜑 (𝑧𝑛)2

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)1/𝑝+2. (39) It follows that

󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑓𝑛󵄩󵄩󵄩󵄩󵄩≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑓𝑛󵄩󵄩󵄩󵄩󵄩Z𝛼

≥ (1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝑢󸀠(𝑧𝑛) 𝜑󸀠(𝑧𝑛) + 𝜑󸀠󸀠(𝑧𝑛) 𝑢 (𝑧𝑛))

× 𝑓𝑛󸀠(𝜑 (𝑧𝑛)) + 𝑢 (𝑧𝑛) 𝑓𝑛󸀠󸀠(𝜑 (𝑧𝑛)) (𝜑󸀠(𝑧𝑛))2 + 𝑢󸀠󸀠(𝑧𝑛) 𝑓𝑛(𝜑 (𝑧𝑛))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨

(𝜑󸀠(𝑧𝑛))2𝑢 (𝑧𝑛) 2𝜑 (𝑧𝑛)2 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)1/𝑝+2

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨 .

(40) Then

𝑛 → ∞lim

(1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧𝑛) (𝜑󸀠(𝑧𝑛))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0. (41) Next, let

𝑔𝑛(𝑧) = − 3 1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2

(1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+1 + 5 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)2 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+2

− 2 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)3 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+3.

(42)

(7)

By a direct calculation we obtain that𝑔𝑛 󴁂󴀱 0 (𝑛 → ∞) on compact subsets of𝐷 and sup𝑛‖𝑔𝑛𝑝 ≤ 𝐶(𝑝) < +∞.

Consequently, {𝑔𝑛} is a bounded sequence in 𝐻𝑝 which converges to 0 uniformly on compact subsets of 𝐷. Then lim𝑛 → ∞‖𝑢𝐶𝜑(𝑔𝑛)‖ = 0by Lemma6. Note that𝑔𝑛(𝜑(𝑧𝑛)) ≡ 0,𝑔󸀠󸀠𝑛(𝜑(𝑧𝑛)) ≡ 0and𝑔𝑛󸀠(𝜑(𝑧𝑛)) = 𝜑(𝑧𝑛)/(1 − |𝜑(𝑧𝑛)|2)1/𝑝+1; it follows that

󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑔𝑛󵄩󵄩󵄩󵄩󵄩 ≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑔𝑛󵄩󵄩󵄩󵄩󵄩Z𝛼

≥ (1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝑢󸀠(𝑧𝑛) 𝜑󸀠(𝑧𝑛) + 𝜑󸀠󸀠(𝑧𝑛) 𝑢 (𝑧𝑛))

× 𝑔󸀠𝑛(𝜑 (𝑧𝑛)) + 𝑢 (𝑧𝑛) 𝑔󸀠󸀠𝑛 (𝜑 (𝑧𝑛)) (𝜑󸀠(𝑧𝑛))2 + 𝑢󸀠󸀠(𝑧𝑛) 𝑔𝑛(𝜑 (𝑧𝑛))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝑢󸀠(𝑧𝑛) 𝜑󸀠(𝑧𝑛) + 𝜑󸀠󸀠(𝑧𝑛) 𝑢 (𝑧𝑛))

× 𝜑 (𝑧𝑛)

(1 −󵄨󵄨󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨󵄨󵄨2)1/𝑝+1

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨

󵄨󵄨󵄨󵄨󵄨󵄨󵄨󵄨.

(43)

Then lim𝑛 → ∞(1 − |𝑧𝑛|2)𝛼(2𝑢󸀠(𝑧𝑛)𝜑󸀠(𝑧𝑛) + 𝜑󸀠󸀠(𝑧𝑛)𝑢(𝑧𝑛))/(1 −

|𝜑(𝑧𝑛)|2)1/𝑝+1= 0.

Finally, let

𝑛(𝑧) = −𝑝 + 3

𝑝 + 1 1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+1

+2 (𝑝 + 3)

𝑝 + 2 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)2 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+2

− (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)3 (1 − 𝜑 (𝑧𝑛)𝑧)1/𝑝+3.

(44)

By a direct calculation we obtain thatℎ𝑛 󴁂󴀱 0 (𝑛 → ∞)on compact subsets of𝐷 and sup𝑛‖ℎ𝑛𝑝 < ∞. Consequently, {ℎ𝑛} is a bounded sequence in 𝐻𝑝 which converges to 0 uniformly on compact subsets of 𝐷. Then lim𝑛 → ∞‖𝑢𝐶𝜑(ℎ𝑛)‖ = 0 by Lemma 6. Note that

𝑛(𝜑(𝑧𝑛)) = 2/(𝑝 + 1)(𝑝 + 2)(1 − |𝜑(𝑧𝑛)|2)1/𝑝,ℎ󸀠𝑛(𝜑(𝑧𝑛)) ≡ 0, andℎ󸀠󸀠𝑛(𝜑(𝑧𝑛)) ≡ 0; it follows that

󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑛󵄩󵄩󵄩󵄩󵄩 ≥ 󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑛󵄩󵄩󵄩󵄩󵄩Z𝛼

≥ (1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝑢󸀠(𝑧𝑛) 𝜑󸀠(𝑧𝑛) + 𝜑󸀠󸀠(𝑧𝑛) 𝑢 (𝑧𝑛))

× ℎ󸀠𝑛(𝜑 (𝑧𝑛))+𝑢 (𝑧𝑛) ℎ󸀠󸀠𝑛 (𝜑 (𝑧𝑛))(𝜑󸀠(𝑧𝑛))2 + 𝑢󸀠󸀠(𝑧𝑛) ℎ𝑛(𝜑 (𝑧𝑛))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

= 2(1 − 󵄨󵄨󵄨󵄨𝑧𝑛󵄨󵄨󵄨󵄨2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧𝑛)󵄨󵄨󵄨󵄨󵄨

(𝑝 + 1) (𝑝 + 2) (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧𝑛)󵄨󵄨󵄨󵄨2)1/𝑝.

(45)

Then lim𝑛 → ∞((1 − |𝑧𝑛|2)𝛼|𝑢󸀠󸀠(𝑧𝑛)|/(1 − |𝜑(𝑧𝑛)|2)1/𝑝) = 0. The proof of the necessary is completed.

Conversely, Suppose that (37) hold. Since 𝑢𝐶𝜑 is a bounded operator, from Theorem4, we have

𝑀1≜sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨 < ∞, 𝑀2≜sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 < ∞, 𝑀3≜sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 < ∞.

(46)

Let{𝑓𝑛}be a bounded sequence in𝐻𝑝with‖𝑓𝑛𝑝 ≤ 1and 𝑓𝑛 → 0uniformly on compact subsets of𝐷. We only prove lim𝑛 → ∞‖𝑢𝐶𝜑(𝑓𝑛)‖ = 0by Lemma6. By the assumption, for any𝜖 > 0, there is a constant𝛿,0 < 𝛿 < 1, such that 𝛿 < |𝜑(𝑧)| < 1implies

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 < 𝜖, (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 < 𝜖, (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 < 𝜖.

(47)

(8)

Let𝐾 = {𝑤 ∈ 𝐷 : |𝑤| ≤ 𝛿}. Note that𝐾is a compact subset of𝐷. Then from Lemma2it follows that

󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑓𝑛󵄩󵄩󵄩󵄩󵄩Z𝛼 =sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑓𝑛)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

= (1 − |𝑧|2)𝛼

×󵄨󵄨󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓𝑛󸀠(𝜑 (𝑧)) + 𝑓𝑛󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)

+ 𝑢󸀠󸀠(𝑧) 𝑓𝑛(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨󵄨󵄨

≤sup

𝑧∈𝐷(1 − |𝑧|2)𝛼

× 󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓𝑛󸀠(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨 +sup

𝑧∈𝐷(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑓𝑛󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

+sup

𝑧∈𝐷(1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧) 𝑓𝑛(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨

≤ 3𝜖 + sup

|𝜑(𝑧)|≤𝛿(1 − |𝑧|2)𝛼

× 󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓𝑛󸀠(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨

+ sup

|𝜑(𝑧)|≤𝛿(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑓𝑛󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

+ sup

|𝜑(𝑧)|≤𝛿(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧) 𝑓𝑛(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨

≤ 3𝜖 + 𝑀3sup

𝑤∈𝐾󵄨󵄨󵄨󵄨󵄨𝑓𝑛󸀠(𝑤)󵄨󵄨󵄨󵄨󵄨 + 𝑀1sup

𝑤∈𝐾󵄨󵄨󵄨󵄨󵄨𝑓𝑛󸀠󸀠(𝑤)󵄨󵄨󵄨󵄨󵄨 + 𝑀2sup

𝑤∈𝐾󵄨󵄨󵄨󵄨𝑓𝑛(𝑤)󵄨󵄨󵄨󵄨 .

(48)

As𝑛 → ∞,

󵄩󵄩󵄩󵄩󵄩𝑢𝐶𝜑𝑓𝑛󵄩󵄩󵄩󵄩󵄩Z𝛼󳨀→ 0. (49) Hence𝑢𝐶𝜑 is compact. This completes the proof of Theo- rem7.

In order to prove the compactness of 𝑢𝐶𝜑 on the little Zygmund spacesZ𝛼,0, we require the following lemma.

Lemma 8. Let𝑈 ⊂Z𝛼,0. Then𝑈is compact if and only if it is closed, bounded and satisfies

|𝑧| → 1limsup

𝑓∈𝑈(1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑓󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 = 0. (50) The proof is similar to that of Lemma 1 in [1], but we omit it.

Theorem 9. Let𝛼 > 0,0 < 𝑝 < ∞,𝑢be an analytic function on the unit disc𝐷and𝜑an analytic self-map of𝐷. Then𝑢𝐶𝜑

is compact from𝐻𝑝to the little Zygmund type spacesZ𝛼,0if and only if (28),(29), and(30)hold.

Proof. Assume that (28), (29), and (30) hold. By Theorem5, we know that𝑢𝐶𝜑is bounded from𝐻𝑝to the little Zygmund type spacesZ𝛼,0. Suppose that𝑓 ∈ 𝐻𝑝with‖𝑓‖𝑝 ≤ 1. From Lemmas1and2we obtain that

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

≤ (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨(2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)) 𝑓󸀠(𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨 + (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑓󸀠󸀠(𝜑 (𝑧)) (𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

+ (1 − |𝑧|2) 󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧) 𝑓 (𝜑 (𝑧))󵄨󵄨󵄨󵄨󵄨

≤ (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

× 1

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝

+(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝,

(51) thus

sup{󵄨󵄨󵄨󵄨󵄨󵄨(1 − |𝑧|2)𝛼(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨 : 𝑓 ∈ 𝐻𝑝, 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝≤ 1}

≤ (1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 1 1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2 +(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 +(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 ,

(52) and it follows that

|𝑧| → 1limsup{󵄨󵄨󵄨󵄨󵄨󵄨(1 − |𝑧|2)𝛼(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨 : 𝑓 ∈ 𝐻𝑝, 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝≤ 1}

= 0.

(53) Hence𝑢𝐶𝜑: 𝐻𝑝 → Z𝛼,0is compact by Lemma8.

Conversely, suppose that𝑢𝐶𝜑: 𝐻𝑝 → Z𝛼,0is compact.

Firstly, it is obvious𝑢𝐶𝜑is bounded, from Theorem5we have𝑢 ∈ Z𝛼,0, and (31), (32) hold. On the other hand, we have

(9)

|𝑧| → 1limsup{󵄨󵄨󵄨󵄨󵄨󵄨(1 − |𝑧|2)𝛼(𝑢𝐶𝜑𝑓)󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨󵄨 : 𝑓 ∈ 𝐻𝑝, 󵄩󵄩󵄩󵄩𝑓󵄩󵄩󵄩󵄩𝑝≤ 𝑀}

= 0,

(54) for some𝑀 > 0by Lemma6.

Next, note that the proof of Theorem4and the fact that the functions given in (18) are in𝐻𝑝and have norms bounded independently of𝑎; we obtain that

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0 (55) for|𝜑(𝑧)| > 1/2. However, if|𝜑(𝑧)| ≤ 1/2, by (31), we easily have

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2

≤ lim

|𝑧| → 1(4

3)1/𝑝+2(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨𝑢 (𝑧) (𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

= 0.

(56)

Similarly, note that the functions given in (22) and (25) are in𝐻𝑝 and have norms bounded independently of𝑎, we obtain that

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 = 0,

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝 = 0

(57)

for|𝜑(𝑧)| > 1/2. However, if|𝜑(𝑧)| ≤ 1/2, from𝑢 ∈Z𝛼,0and (32), we easily have

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1

≤ (4

3)1/𝑝+1 lim

|𝑧| → 1(1 − |𝑧|2)𝛼

× 󵄨󵄨󵄨󵄨󵄨2𝜑󸀠(𝑧) 𝑢󸀠(𝑧) + 𝜑󸀠󸀠(𝑧) 𝑢 (𝑧)󵄨󵄨󵄨󵄨󵄨

= 0,

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝

≤ lim

|𝑧| → 1(4

3)1/𝑝(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝑢󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨

= 0.

(58)

This completes the proof of Theorem9.

Remark 10. From Theorems5and9, we conjecture that𝑢𝐶𝜑: 𝐻𝑝 → Z𝛼,0is compact if and only if𝑢𝐶𝜑 : 𝐻𝑝 → Z𝛼,0is bounded.

Taking 𝑢(𝑧) = 1 from Theorems 4, 7, and 9, we obtain the following results about the characterization of the boundedness and compactness of the composition operator 𝐶𝜑: 𝐻𝑝 → Z𝛼(orZ𝛼,0).

Corollary 11. Let𝛼 > 0,0 < 𝑝 < ∞, and𝜑be an analytic self-map of𝐷. Then𝐶𝜑 : 𝐻𝑝 → Z𝛼is a bounded operator if and only if the following are satisfied:

sup𝑧∈𝐷

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 < ∞, sup𝑧∈𝐷

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝜑󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 < ∞.

(59)

Corollary 12. Let𝛼 > 0,0 < 𝑝 < ∞, and𝜑be an analytic self-map of𝐷. Then𝐶𝜑 : 𝐻𝑝 → Z𝛼is a compact operator if and only if𝐶𝜑is bounded and the following are satisfied:

|𝜑(𝑧)| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0,

|𝜑(𝑧)| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝜑󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 = 0.

(60)

Corollary 13. Let𝛼 > 0,0 < 𝑝 < ∞, and𝜑be an analytic self-map of𝐷. Then𝐶𝜑 : 𝐻𝑝 → Z𝛼,0is a compact operator if and only if

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨󵄨(𝜑󸀠(𝑧))2󵄨󵄨󵄨󵄨󵄨󵄨

(1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+2 = 0,

|𝑧| → 1lim

(1 − |𝑧|2)𝛼󵄨󵄨󵄨󵄨󵄨𝜑󸀠󸀠(𝑧)󵄨󵄨󵄨󵄨󵄨 (1 − 󵄨󵄨󵄨󵄨𝜑 (𝑧)󵄨󵄨󵄨󵄨2)1/𝑝+1 = 0.

(61)

In the formulation of corollary, we use the notation𝑀𝑢on 𝐻(𝐷)defined by𝑀𝑢𝑓 = 𝑢𝑓for𝑓 ∈ 𝐻(𝐷). Taking𝜑(𝑧) = 𝑧 from Theorems4,5,7, and9, we obtain the following results about the characterization of the boundedness and compact- ness of pointwise multiplier𝑀𝑢: 𝐻𝑝 → Z𝛼(𝑜𝑟 Z𝛼,0).

Corollary 14. Let𝛼 > 0,0 < 𝑝 < ∞. Then the pointwise multiplier𝑀𝑢: 𝐻𝑝 → Z𝛼is a bounded operator if and only if

(i)𝑢 = 0if𝛼 < 2 + 1/𝑝;

(ii)𝑢 ∈ 𝐻if𝛼 = 2 + 1/𝑝;

(iii) sup𝑧∈𝐷(1 − |𝑧|2)𝛼−2−1/𝑝|𝑢(𝑧)| < ∞if𝛼 > 2 + 1/𝑝.

参照

関連したドキュメント

[2] A.D.Gadjiev, A.Aral, The estimates of approximation by using a new type of weighted modulus of

In addition, some properties of the numerical range of bounded linear operators on weighted Hardy spaces are discussed.. 2 weighted

Heinig, Weighted norm inequalities for certain integral operators.. Stepanov, Two-weighted estimates for Riemann–Liouville

It considers two different versions of the Hardy operator and characterizes their weighted weak type inequalities when p = 1.. It proves that for the classical Hardy operator, the

The issues of boundedness and compactness of the composition operators and of the weighted composition operators between different analytic function spaces on the unit disc D , the

With L p , 1 ≤ p &lt; ∞, spaces as our motivation, the completeness and inductive limits of those spaces are studied including their relationship with the weighted spaces of

An. Timi¸ soara Ser. Composition operators on Hardy spaces of a half- plane.. Composition operators and classical function theory. Uni- versitext: Tracts in Mathematics. Hardy

It turns out that the class of weights corresponding to I α acting on weak−L φ for φ of lower type equal or greater than n/α, is the same as the one solving the problem for weak−L