00060000713 một vài lớp toán tử trên các không gian hàm 00060000713 một vài lớp toán tử trên các không gian hàm
Trang 1ĐẠI HỌC QUỐC GIA HÀ NỘI
KẾT QUẢ THỰC HIỆN ĐỀ TÀI KH&CN
CẤP ĐẠI HỌC QUỐC GIA
Tên đề tài: Một vài lớp toán tử trên các không gian hàm
Mã số đề tài: QG.19.12
Chủ nhiệm đề tài: TS Phạm Trọng Tiến
Hà Nội, 2020
Trang 21
PHẦN I THÔNG TIN CHUNG
1.1 Tên đề tài: Một vài lớp toán tử trên các không gian hàm
1.2 Mã số: QG.19.12
1.3 Danh sách chủ trì, thành viên tham gia thực hiện đề tài
TT Chức danh, học vị, họ và tên Đơn vị công tác Vai trò thực hiện đề tài
học Tự nhiên
Chủ nhiệm đề tài
Ghi chú: Ngoài ra trong đề tài còn có sự tham gia thực hiện của 01 học viên cao học
1.4 Đơn vị chủ trì:
Tên đơn vị chủ trì: Trường Đại học Khoa học Tự Nhiên
Điện thoại: 0243-8581419 Fax: 0243-8583061
E-mail: hus@vnu.edu.vn
1.5 Thời gian thực hiện:
1.5.1 Theo hợp đồng: từ tháng 12 năm 2018 đến tháng 12 năm 2020
1.5.2 Gia hạn (nếu có): đến tháng… năm…
1.5.3 Thực hiện thực tế: từ tháng 12 năm 2018 đến tháng 12 năm 2020
1.6 Những thay đổi so với thuyết minh ban đầu (nếu có):
(Về mục tiêu, nội dung, phương pháp, kết quả nghiên cứu và tổ chức thực hiện; Nguyên nhân; Ý kiến của Cơ quan quản lý)
Không có thay đổi sơ với thuyết minh ban đầu
1.7 Tổng kinh phí được phê duyệt của đề tài: 320 triệu đồng
PHẦN II TỔNG QUAN KẾT QUẢ NGHIÊN CỨU
1 Đặt vấn đề
Có thể nói trong thời đại bùng nổ về khoa học và công nghệ như hiện nay thì Toán học đóng một vai trò vô cùng quan trọng Hiện nay chúng ta có thể dễ dàng nhận ra rằng rất nhiều vấn đề trong khoa học kỹ thuật, trong kinh tế, và trong xã hội được nghiên cứu, dự báo, và giải quyết thông qua các mô hình toán học khác nhau Nói một cách cụ thể hơn, trong mỗi mô hình toán học được sử dụng sẽ có một số thành phần đặc trưng cho những đại lượng thực tế và những đại lượng đó sẽ thay đổi theo thời gian và không gian Để nghiên cứu đại lượng này, các nhà toán học thường xây dựng các không gian nền đủ tổng quát để mô tả các đối tượng và tìm cách tìm hiểu các tính chất của chúng thông qua việc nghiên cứu sự tương tác của chúng với các lớp toán tử khác nhau Việc xây dựng các lớp
Trang 32
trong thực tiễn Với lý do trên, các thành viên tham gia đề tài này muốn nghiên cứu một
số tính chất và đưa ra các ước lượng cho một số lớp toán tử trên các không gian hàm khác nhau
2 Mục tiêu
Kí hiệu X là không gian các hàm chỉnh hình trên miền G trong mặt phẳng phức C Với
hai hàm chỉnh hình : G G và : G C , toán tử kết hợp có trọng W , được định nghĩa như sau:
,
W ( ) : f f , f X Khi hàm bằng đồng nhất 1 trên G, thì toán tử W , sẽ trở thành toán tử kết hợp C Vấn đề trọng tâm trong nghiên cứu hai toán tử này là tìm hiểu mối liên hệ giữa các tích chất của các hàm và với các tính chất của các toán tử W , và C Những toán tử này đã thu hút được sự quan tâm của các nhà toán học từ những năm 80 của thế kỷ trước
và được nghiên cứu trên những không gian hàm khác nhau như: không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Bloch, không gian Fock; và theo nhiều hướng khác nhau như: tính liên tục [1, 2, 3, 4], tính chất phổ [4, 5, 6], hiệu compact [7], cấu trúc tôpô [8, 9], tính chất động lực học và tính ergodic [10, 11],
Tuy vậy, trên một số không gian đặc biệt thì những vấn đề này lại là rất khó và vẫn chưa được giải quyết hoàn thiện
Bên cạnh hai toán tử trên, toán tử vi phân và tích phân cũng là một lớp toán tử cơ bản
và được nghiên cứu nhiều trên các không gian hàm khác nhau, đặc biệt là các ước lượng cho các toán tử vi phân và tích phân đóng vai trò vô cùng quan trọng trong giải tích nói chung, giải tích hàm và phương trình đạo hàm riêng nói riêng Trong các tài liệu [12, 13, 14] các tác giả đã đưa ra một số đánh giá giữa độ tăng của đạo hàm và độ tăng của hàm nguyên thông qua công thức Littlewood-Paley, từ đó các tác giả đã đặc trưng được tính bị chặn cho toán tử tích phân Volterra suy rộng
Trong các tài liệu [15, 16] các tác giả đã nghiên cứu dạng ngược của bất đẳng thức Hardy-Littewood-Sobolev (HLS) cho một lớp các hàm khả tích Bất đẳng thức HLS là một dạng suy rộng của bất đẳng thức Young nổi tiếng trong tích chập cho các hàm khả tích mà ở đó một trong số các nhân có chứa điểm kỳ dị trên không gian lấy tích phân Kết quả thu được ở [15, 16] cho các trường hợp Rn và nửa không gian Rn+ là quan trọng vì cùng với dạng thuận của bất đẳng thức HLS nó cung cấp các ước lượng liên quan đến các toán tử tích phân với nhân kì dị Trong giải tích từ rất lâu người ta đã biết các ước lượng cho các toán tử tích phân với nhân kì dị đóng vai trò rất quan trọng Việc chứng minh được dạng ngược của bất đẳng thức HLS đã mở ra hướng nghiên cứu mới thu hút các nhà toán học trên thế giới mà ở đó một lớp rộng hơn các toán tử tích phân với nhân kỳ dị được nghiên cứu, chẳng hạn trong tác phẩm [17] được thực hiện bởi các chuyên gia hàng đầu về giải tích và phương trình đạo hàm riêng
Dựa trên những đánh giá tổng quan ở trên, trong đề tài này các thành viên dự định nghiên cứu những nội dung chính sau đây:
Vấn đề 1: Nghiên cứu tính liên tục và tính compact cho toán tử kết hợp C và toán
tử kết hợp có trọng W , trên các không gian hàm khác nhau Những nghiên cứu này sẽ là cơ sở cho những nghiên cứu tiếp theo đối với hai toán tử này
Trang 43
Vấn đề 2: Nghiên cứu tính chất động lực học và tính ergodic cho toán tử kết hợp
C và toán tử kết hợp có trọng W , trên các không gian hàm khác nhau Đây là những hướng nghiên cứu rất thời sự và có nhiều ý nghĩa trong các ngành kỹ thuật Bài toán quan trọng trong hướng nghiên cứu này là tìm hiểu dáng điệu của dãy lặp các toán tử, từ đó có thể đưa ra những dự đoán đối với các bài toán trong thực tế
Vấn đề 3: Nghiên cứu các tính chất của một số lớp toán tử tích phân và vi phân trên
một số không gian hàm thông dụng dựa vào việc tìm ra các ước lượng điểm hoặc ước lượng tích phân Chúng tôi sẽ tập trung thực hiện các nghiên cứu liên quan đến các bất đẳng thức hàm trên các không gian mà chúng tôi quan tâm, chẳng hạn, không gian Fock, không gian Sobolev, không gian Fock-Sobolev, nhóm Heisenberg Một phần của Đề tài liên quan đến vấn đề này là đi tìm câu trả lời cho một số câu hỏi mở
Tài liệu tham khảo:
[1] M D Contreras, A G Hernández-Díaz, Weighted composition operators between
different Hardy spaces, Integr Equa Oper Theory 46 (2003), 871–887
[2] Z Cuckovic, R Zhao, Weighted composition operators on the Bergman space, J London Math Soc (2) 70 (2004), 499–511
[3] T Le, Normal and isometric weighted composition operators on the Fock space,
Bull Lond Math Soc 46 (2014), 847–856
[4] I Chalendar, E A Gallardo-Gutierrez, J R Partington, Weighted composition
operators on the Dirichlet space: boundedness and spectral properties, Math Ann
363 (2015), 1265–1279
[5] R Aron, M Lindstrom, Spectra of weighted composition operators on weighted
Banach spaces of analytic functions, Israel J Math 141 (2004), 263–276
[6] S Hyvarinen, M Lindstrom, I Nieminen and E Saukko, Spectra of weighted
composition operators with automorphic symbols, J Funct Anal 265 (2013), no 8,
1749–1777
[7] J Moorhouse, Compact differences of composition operators, J Funct Anal 219
(2005), 70–92
[8] E A Gallardo-Gutierrez, M J Gonzalez, P J Nieminen, E Saksman, On the
connected component of compact composition operators on the Hardy space, Adv Math 219 (2008), 986–1001
[9] K J Izuchi, S Ohno, Path connected components in weighted composition
operators on h1 and H1 with the operator norm, Trans Amer Math Soc 365 (2013),
3593–3612
[10] B Yousefi, H Rezaei, Hypercyclic property of weighted composition operators,
Proc Amer Math Soc 135 (2007), 3263–3271
[11] M J Beltrán-Meneu, M C Gómez-Collado, E Jordá, D Jornet, Mean ergodic
composition operators on Banach spaces of holomorphic functions, J Funct Anal
270 (2016), 4369–4385
[12] O Constantin, A Volterra-type integration operator on Fock spaces, Proc Amer Math Soc 140 (2012) 4247–4257
[13] O Constantin, J A Peláez, Integral operators, embedding theorems and a
Littlewood-Paley formular on weighted Fock spaces, J Geom Anal 26 (2015)
1109-1154
Trang 54
operators, Proc Amer Math Soc 141 (2013), 2829–2840
[15] Quoc Anh Ngo, V.H Nguyen, Sharp reversed Hardy-Littlewood-Sobolev inequality
on Rn, Israel Journal of Mathematics 220 (2017) 189–223
[16] Quoc Anh Ngo, V.H Nguyen, Sharp reversed Hardy-Littlewood-Sobolev inequality
on the half space Rn+, International Mathematics Research Notices 2017 (2017)
6187–6230
[17] J.A Carrillo, M.G Delgadino, J Dolbeault, R.L Frank, F Hoffmann, Reverse Hardy-Littlewood-Sobolev inequalities, arXiv:1807.09189
3 Phương pháp nghiên cứu
Đối với vấn đề 1: Chúng tôi sẽ sử dụng những kiến thức cơ bản của Giải tích phức, Giải tích hàm, Lý thuyết toán tử và những kết quả nghiên cứu về tính chất của các không gian hàm đã có
Đối với vấn đề 2: Chúng tôi sẽ sử dụng những kiến thức và kỹ thuật trong Lý thuyết động lực học và Lý thuyết ergodic, kết hợp với những kết quả nghiên cứu về tính liên tục
và tính compact cho toán tử kết hợp và toán tử kết hợp có trọng trong vấn đề 1
Đối với vấn đề 3: Chúng tôi sẽ sử dụng những kiến thức và kỹ thuật trong Giải tích thực, Giải tích hàm, Giải tích điều hoà, và Lý thuyết định tính phương trình đạo hàm riêng
Các kỹ thuật và cách tiếp cận mà tác giả dự định triển khai trong đề tài đều là mới và chứa đựng nhiều đánh giá, ước lượng và kỹ thuật tinh tế
Phương pháp nghiên cứu trong đề tài chứa đựng nhiều ý tưởng mới cho những vấn đề tiếp theo trong những hướng nghiên cứu này
4 Tổng kết kết quả nghiên cứu
Vấn đề 1: Bản tổng quan các kết quả về tính bị chặn của toán tử kết hợp C trên các không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Banach có trọng các hàm chỉnh hình Đặc biệt, các kết quả về tính bị chặn và tính compact cho toán
tử W , và C giữa các không gian Fock
Vấn đề 2: Các kết quả về tính hội tụ đều, hội tụ mạnh, hội tụ yếu, và tính chất
ergodic cho dãy lặp của toán tử kết hợp C trên không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Banach có trọng các hàm chỉnh hình, không gian Bloch
có trọng
Vấn đề 3: Đưa ra được dạng tốt nhất của bất đẳng thức Sobolev vết bậc cao trên
quả cầu Bn trong Rn và bất đẳng thức Sobolev bậc cao trên tới hạn cho các hàm đối xứng cầu trên Bn
trong Rn
5 Đánh giá về các kết quả đã đạt được và kết luận
Vấn đề 1: Các kết quả về tính bị chặn và tính compact của toán tử kết hợp C và toán tử kết hợp có trọng W , trên các không gian khác nhau là cơ sở cho những nghiên cứu xa hơn về hai toán tử W , và C
Trang 65
Vấn đề 2: Các kết quả về dãy lặp của toán tử kết hợp C là mới, có tính thời sự cao, và thu hút được sự quan tâm của các nhà toán học Những kết quả này còn mở ra những ý tưởng để cho nghiên cứu dãy lặp của toán tử kết hợp có trọng W ,
Vấn đề 3: Các kết quả về bất đẳng thức Sobolev trên quả cầu Bn là mới, có tính thời sự cao, và thu hút được sự quan tâm trong cộng đồng, thể hiện qua số lần được trích dẫn bởi các tác giả khác Các kết quả này là công cụ hữu ích khi nghiên cứu các bài toán biên trên quả cầu Bn
và mở ra một vấn đề là phát triển các bất đẳng thức này trên không
gian Fock và Fock-Sobolev Bài toán này sẽ tiếp tục được nghiên cứu trong tương lai gần
6 Tóm tắt kết quả (tiếng Việt và tiếng Anh)
Vấn đề 1: Các đặc trưng của tính bị chặn của toán tử kết hợp C trên các không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Banach có trọng các hàm chỉnh hình đã được tổng hợp và trình bày lại Đặc biệt, điều kiện cần và đủ cho tính
bị chặn và tính compact của hai toán tử W , và C giữa các không gian Fock có trọng đã được thiết lập
Vấn đề 2: Tính hội tụ theo chuẩn, hội tụ mạnh, hội tụ yếu, và các tính chất ergodic
của dãy lặp của toán tử kết hợp Ctrên không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Banach có trọng các hàm chỉnh hình, không gian Bloch có trọng đã được nghiên cứu đầy đủ Cụ thể, trên những không gian nêu trên những kết quả sau đây đã được chứng minh:
Nếu hàm là một đẳng cấu elliptic (elliptic automorphism) thì dãy lặp của toán tử Ckhông hội tụ yếu và C là mean ergodic
Nếu hàm có điểm Denjoy-Wolff nằm trong hình cầu đơn vị D thì dãy lặp của toán tử C hội tụ đều
Nếu hàm có điểm Denjoy-Wolff nằm trên biên của hình cầu đơn vị D thì dãy lặp của toán tử C không là mean ergodic
Vấn đề 3: Đối với bất đẳng thức Sobolev vết bậc cao trên quả cầu Bn trong Rn, kết quả thu được là các đánh giá tối ưu cho chuẩn Lp
của vết của hàm trên mặt cầu Sn-1 thông qua chuẩn của hàm trong không gian Sobolev thích hợp trên quả cầu Bn Đối với bất đẳng thức Sobolev bậc cao trên tới hạn cho các hàm đối xứng cầu trên Bn, kết quả thu được chỉ
ra rằng lớp các hàm đối xứng cầu thuộc không gian Sobolev Hm có thể được nhúng vào không gian Lp
trên Bn với p trên tới hạn nào đó Để minh họa kết quả, một ví dụ áp dụng
về sự tồn tại nghiệm của một lớp các bài toán biên cho toán tử đa điều hòa Δm
đã được
minh họa
Problem 1: The well-known characterizations for the boundedness of composition
operators on Hardy spaces, Bergman spaces, Dirichlet spaces, weighted Banach spaces with sup-norm are surveyed Moreover, necessary and sufficient conditions for the boundedness and compactness of composition operators and weighted composition operators between Fock spaces are established
Problem 2: The uniform convergence, strong convergence, weak convergence, and
ergodicity of the iterates of composition operators C are studied on various Banach spaces of holomorphic functions on the unit disc, such as Bergman spaces, Dirichlet
Trang 76
spaces, the following results are proved:
The iterates of the operator C do not converge weakly and C is mean ergodic if is an elliptic automorphism
The iterates of C converge uniformly if the Denjoy-Wolff point of is in the unit disc D
C is not mean ergodic if the Denjoy-Wolff point of lies on the boundary
of the unit disc D
Problem 3: For the higher order Sobolev trace inequality on Bn in Rn, our finding consists of sharp estimates of Lp norm of the trace of functions on the sphere Sn-1 in terms
of the norm of them in a suitable Sobolev space on the ball Bn For the supercritical higher order Sobolev inequality for radially symmetric functions on Bn, our finding indicates that the set of radially symmetric functions in the Sobolev Hm can be embedded into an Lpspace on Bn for some supercritical p To illustrate the finding, an application to the existence of solutions for a class of boundary value problems for the polyharmonic operator Δm is provided
PHẦN III SẢN PHẨM, CÔNG BỐ VÀ KẾT QUẢ ĐÀO TẠO CỦA ĐỀ TÀI
3.1 Kết quả nghiên cứu
TT Tên sản phẩm Yêu cầu khoa học hoặc/và chỉ tiêu kinh tế - kỹ thuật
trọng trên các không gian
không gian Hardy, không
gian Bergman, không gian
Dirichlet
- Điều kiện cần và đủ cho tính liên tục và tính compact của hai toán tử
,
W và C phải được thiết lập theo hai hàm
và
- Đặc trưng cho tính động lực học và tính ergodic của hai toán tử W , và
C phải được thiết lập theo hai hàm và
- Điều kiện cần và đủ cho tính liên tục và tính compact của hai toán tử
,
W và C phải được thiết lập theo hai hàm
và
- Đặc trưng cho tính động lực học và tính ergodic của hai toán tử W , và
C phải được thiết lập theo hai hàm và
Trang 87
2
lượng liên quan đến các
toán tử vi phân trên không
gian Fock, không gian
Sobolev, không gian
Fock-Sobolev
- Các đánh giá và ước lượng được thể hiện thông qua một số bất đẳng thức hàm trên một số không gian hàm
- Các đánh giá và ước lượng được thể hiện thông qua một số bất đẳng thức hàm trên một số không gian hàm
ơn sự tài trợ của ĐHQGHN đúng quy định
Đánh giá chung
(Đạt, không đạt)
1 Công trình công bố trên tạp chí khoa học quốc tế theo hệ thống
ISI/Scopus
1.1 Pham Trong Tien,
The iterates of composition operators on
Banach spaces of holomorphic functions,
Journal of Mathematical Analysis and
1.3 Quoc Anh Ngo, Van Hoang Nguyen,
A supercritical Sobolev type inequality in
higher order Sobolev spaces and related
higher order elliptic problems,
Journal of Differential Equations, 268
Trang 97 Kết quả dự kiến được ứng dụng tại các cơ quan hoạch định chính sách
hoặc cơ sở ứng dụng KH&CN
7.1
Ghi chú:
- Cột sản phẩm khoa học công nghệ: Liệt kê các thông tin các sản phẩm KHCN theo thứ tự <tên tác giả, tên công trình, tên tạp chí/nhà xuất bản, số phát hành, năm phát hành, trang đăng công trình, mã công trình đăng tạp chí/sách chuyên khảo (DOI), loại tạp chí ISI/Scopus>
- Các ấn phẩm khoa học (bài báo, báo cáo KH, sách chuyên khảo…) chỉ được chấp nhận nếu có ghi nhận địa chỉ và cảm ơn tài trợ của ĐHQGHN theo đúng quy định
- Bản phô tô toàn văn các ấn phẩm này phải đưa vào phụ lục các minh chứng của báo cáo Riêng sách chuyên khảo cần có bản phô tô bìa, trang đầu và trang cuối có ghi thông tin mã số xuất bản
3.3 Kết quả đào tạo
29/12/2019, nhận bằng ngày 22/05/2020
Ghi chú:
- Gửi kèm bản photo trang bìa luận án/ luận văn/ khóa luận và bằng hoặc giấy chứng nhận nghiên cứu sinh/thạc sỹ nếu học viên đã bảo vệ thành công luận án/ luận văn;
- Cột công trình công bố ghi như mục III.1
PHẦN IV TỔNG HỢP KẾT QUẢ CÁC SẢN PHẨM KH&CN VÀ ĐÀO TẠO CỦA
ĐỀ TÀI
lượng đăng ký
Số lượng
đã hoàn thành
Trang 1003 bài ISI
2 Sách chuyên khảo được xuất bản hoặc ký hợp đồng
xuất bản
5 Số lượng bài báo trên các tạp chí khoa học của
ĐHQGHN, tạp chí khoa học chuyên ngành quốc gia
hoặc báo cáo khoa học đăng trong kỷ yếu hội nghị
7 Kết quả dự kiến được ứng dụng tại các cơ quan hoạch
định chính sách hoặc cơ sở ứng dụng KH&CN
(triệu
đồng)
Ghi chú
2 Nguyên, nhiên vật liệu, cây con
và viết tổng quan tài liệu
Trang 1110
chức thực hiện ở các cấp)
Không có
PHẦN VI PHỤ LỤC (minh chứng các sản phẩm nêu ở Phần III)
6.1 Báo cáo sản phẩm khoa học
TT Tên sản phẩm Yêu cầu khoa học hoặc/và chỉ tiêu kinh tế - kỹ thuật Ghi chú
gian Hardy, không gian
Bergman, không gian
không gian Bergman,
không gian Dirichlet
- Điều kiện cần và đủ cho tính liên tục và tính compact của toán tử kết hợp theo điều kiện của không gian
và hàm sinh
- Đặc trưng cho tính động lực học và tính ergodic của của toán tử kết hợp theo điều kiện của không gian và
hàm sinh
Kết quả được thể hiện trong 01 bài báo ISI của Phạm Trọng Tiến và trong luận văn tốt nghiệp của học viên cao học
2 - Một số đánh giá và
ước lượng liên quan đến
các toán tử vi phân trên
không gian Fock, không
gian Sobolev, không
gian Fock-Sobolev
- Các đánh giá và ước lượng được thể hiện thông qua một số bất đẳng thức hàm trên một số không gian hàm
Kết quả được thể hiện trong 02 bài báo ISI của Ngô Quốc Anh và đồng tác giả
6.2 Hình thức và cấp độ công bố kết quả
Các kết quả được thể hiện trong 03 bài báo quốc tế thuộc danh mục ISI:
1 Pham Trong Tien, The iterates of composition operators on Banach spaces of
holomorphic functions, Journal of Mathematical Analysis and Applications, 487
(2020), 123945 (ISI)
DOI: 10.1016/j.jmaa.2020.123945
Trang 12Thơm Phạm Trọng Tiến Luận văn thạc sĩ với đề tài “Toán tử kết hợp có trọng giữa
không gian Fock F(C ) và
)
(C
Đã bảo vệ ngày
29/12/2019, nhận bằng ngày
Trang 1312
KẾT HỢP VÀ TOÁN TỬ KẾT HỢP CÓ TRỌNG
I Nội dung
Cho X là không gian các hàm chỉnh hình trên miền G trong mặt phẳng phức C Với hai
hàm chỉnh hình : G G và : G C , toán tử kết hợp có trọng W , được định nghĩa như sau: W , ( ) : f f , f X Khi hàm bằng đồng nhất 1 trên G, thì toán tử W ,
sẽ trở thành toán tử kết hợp C Vấn đề trọng tâm trong nghiên cứu hai toán tử này là tìm hiểu mối liên hệ giữa các tích chất của các hàm và với các tính chất của các toán tử
,
W và C Tính liên tục và tính compact của hai toán tử này đã được nghiên cứu nhiều trên các không gian hàm chỉnh hình trên đĩa đơn vị D Các kết quả về tính liên tục của toán tử kết hợp C trên không gian Hardy, không gian Bergman, không gian Diriclet, không gian Banach có trọng các hàm chỉnh hình, không gian Bloch có trọng đã được hệ thống lại trong công trình [1] bên dưới Trên không gian Fock gồm các hàm chỉnh hình trên toàn bộ mặt phẳng phức C, các kết quả về tính bị chặn và tính compact được thể hiện trong các định lý sau đây
Định lí 1: Cho p ( 0 , ) , là một hàm khác 0 trong Fp(C ) và ( z ) az b với
(
| ) (
| ) , (
z z
1 Pham Trong Tien, The iterates of composition operators on Banach spaces of
holomorphic functions, Journal of Mathematical Analysis and Applications, 487
Trang 14Trong bài báo [1], chúng tôi đã nghiên cứu đầy đủ những tính chất nêu trên cho dãy lặp của toán tử kết hợp Ctrên không gian Hardy, không gian Bergman, không gian Dirichlet, không gian Banach có trọng các hàm chỉnh hình, không gian Bloch có trọng Cụ thể, trên những không gian nêu trên, những kết quả sau đây đã được chứng minh:
Nếu hàm là một đẳng cấu elliptic (elliptic automorphism) thì dãy lặp của toán tử Ckhông hội tụ yếu và toán tử C là mean ergodic
Nếu hàm có điểm Denjoy-Wolff nằm trong hình cầu đơn vị D thì dãy lặp của toán tử C hội tụ đều
Nếu hàm có điểm Denjoy-Wolff nằm trên biên của hình cầu đơn vị D thì toán tử C không là mean ergodic
II Minh chứng
1 Pham Trong Tien, The iterates of composition operators on Banach spaces of
holomorphic functions, Journal of Mathematical Analysis and Applications, 487
(2020), 123945 (ISI)
DOI: 10.1016/j.jmaa.2020.123945
Trang 1514
I Nội dung
a Bất đẳng thức vết Sobolev trên quả cầu Bn
Năm 2017, Ache và Chang đưa ra dạng tối ưu cho bất đẳng thức vết Sobolev bậc 4 trên quả cầu Bn
sau đây
1
3 2
5 6
7 8
trên quả cầu Bn; xem các Định lý 1.3 và 6.2
b Bất đẳng thức Sobolev trên tới hạn
Định lý nhúng Sobolev cổ điển nói rằng phép nhúng liên tục 1
0( n) p( n)
đúng nếu p không trên tới hạn, tức là p≤2n/(n-2) Năm 2016, các tác giả do Ó, Ruf, và Ubilla chỉ ra rằng nếu hạn chế trên lớp các hàm đối xứng cầu thì phép nhúng trên có thể được cải tiến với p>2n/(n-2) Cụ thể là phép nhúng 1 2 * | |
1 Quoc Anh Ngo, Van Hoang Nguyen, Quoc Hung Phan, Higher order Sobolev trace
inequalities on balls revisited, Journal of Functional Analysis, 278 (2020), 108414
Trang 16Contents lists available at ScienceDirect
Journal of Mathematical Analysis and Applications
www.elsevier.com/locate/jmaa
The iterates of composition operators on Banach spaces of
holomorphic functions
Pham Trong Tien1
Faculty of Mathematics, Mechanics and Informatics, VNU University of Science, Vietnam National
University, Hanoi, Vietnam
Dedicated to Prof Le Hai Khoi on
the occasion of his 60th birthday
(i) the iteratesC n
ϕdo not converge in the weak operator topology andC ϕis mean ergodic ifϕ isan elliptic automorphism,
(ii) C n
ϕconverge uniformly if the Denjoy-Wolff point ofϕ isin D,
(iii) Cϕ is not mean ergodic if the Denjoy-Wolff point of ϕ lies on the boundary
L(X); converges in the strong operator topology (briefly, in the SOT ) if T n x converges in X for every
x ∈ X;andconverges in the weak operator topology (briefly, in the WOT )ifT n x convergesweaklyinX for
everyx ∈ X.AnoperatorT ∈ L(X) iscalled mean ergodic iftheCesàro means
Trang 17converge intheSOT,anduniformly mean ergodic if T [n] convergeuniformly.It iswellknownthatuniformconvergence,strongconvergence,andweakconvergenceareorderedindecreasingstrength;uniformconver-gencealsoimpliesuniformmeanergodicity,whichisstrongerthanmeanergodicity;moreover,ifT is power bounded (i.e.supn T n L(X) < ∞)andthesequenceT n convergesintheWOT,thenT ismeanergodic.Accordingto 20,Theorem1.3] and 23,ChapterVIII,Section3],ifT ispowerbounded,thenT ismeanergodicpreciselywhenX = Ker(I −T ) ⊕Im(I − T );inthiscaseImP = Ker(I −T ) and KerP = Im(I − T ),
where P x := limn→∞ T [n] x, x ∈ X. Moreover, in 15] it was shownthat if T n L(X) /n → 0 as n → ∞,
then T is uniformly mean ergodic ifand only ifIm(I − T ) is closed; in this caseKerP = Im(I − T ). Formoreinformationaboutmeanergodicoperators,see 20,23] andthereferencestherein.InthispaperweareinterestedinthesetypesofconvergenceforcompositionoperatorsondifferentBanachspacesofholomorphicfunctions
Let H(D) denote the space of all holomorphic functions on the unit disk D endowed with the usualcompact open topologyand S(D) theset of allholomorphic self-maps of D For afunction ϕ ∈ S(D),a
composition operator C ϕisdefinedbyC ϕ f := f ◦ ϕ, f ∈ H(D).Compositionoperatorsonvariousspacesofholomorphic functionshave been intensivelyinvestigated duringthe pastfewdecades inmany directions,suchasboundednessandcompactness,essentialnorm,compactdifferences,topologicalstructure(see 10,21]and referencesthereinformoreinformation)
Recently, the study of dynamical and ergodic properties of C ϕ has received a special attention frommanyauthors.TheergodicityofcompositionoperatorswasinvestigatedonthespaceH(U ) ofholomorphicfunctions onadomain U inaStein manifoldby Bonetand Domański 7] and onweightedBanachspaceswithsup-normbyWolf 22].In 5 theauthorscharacterizedmeanergodicityanduniformmeanergodicity
of composition operators on the disk algebra A(D) and on the space H ∞(D) of bounded holomorphicfunctions Later, Arendt, Chalendar, Kumar, and Srivastava 3] studied the asymptotic behavior of thepowers C n
ϕ ondiverse Banachspaces ofholomorphic functions,suchas HardyspacesH p,weightedHardyspacesH p (β),theWieneralgebraW (D), A(D),andH ∞(D).TheauthorsprovedthateitherthepowersC n
ϕ
converge uniformly or theydonotconverge even intheSOT.Inparticular, in 3, Section 5] itwasshownthat on spacesW (D), A(D), and H ∞(D), mean ergodicityalso implies uniform convergence of C n
ϕ whentheDenjoy-Wolffpointofϕ isinD
TheaimofthispaperistocontinuethistopicforcompositionoperatorsonotherclassicalBanachspaces
of holomorphic functions,suchas Bergman spaces A p
auniquepointz0∈ D,calledtheDenjoy-Wolff point,suchthatthesequence ϕ n converges toz0uniformly
oncompactsubsetsofD.Recallthatafunctionϕ ∈ S(D) issaidtobeanelliptic automorphism if ϕ isanautomorphism fixing apoint inD; moreover, anelliptic automorphism ϕ is called of order k, if ϕ k is theidentity forsomesmallestnumberk ∈ N, andof infinite order otherwise.
The paper is organized as follows In Section 2 we recall the definition of classical Banach spaces ofholomorphicfunctionsontheunitdisk and someauxiliaryresultsoncompositionoperators inthem.Section3isdevotedtothesequenceC ϕ ninducedbyanellipticautomorphism.Inthiscaseweprovethat
onaBanachspaceX ofholomorphicfunctionsundersomenaturalassumptions, C ϕ n doesnotconvergeintheWOT,C ϕis alwaysmeanergodic,anditisuniformly meanergodicifandonlyifϕ isoffiniteorder k
Trang 18weprovethatthesequenceC ϕ nalsoconvergesuniformlyonD p
αwithp and α satisfying(4.1),providedthattheoperatorC ϕ ispowerbounded onD p for someγ ∈ (−1, α).Inparticular, foreveryunivalent function
ϕ ∈ S(D) with Denjoy-Wolffpoint inD, this resultholds on Dirichletspaces D p
α with p and α satisfying
(4.2).Moreover,wealsoshowthatC ϕ inducedbyaninnerfunctionϕ isnotuniformlymeanergodicontheHardyspaceH p (Theorem4.2).Thiscomplementstheresultsin 3] forC ϕ onH p
Finally,Section5isrelatedtocompositionoperatorsC ϕinducedbyϕ ∈ S(D) withaboundaryWolffpoint Inthis casethesequence C ϕ n isnot powerbounded andnotmean ergodicon manyclassicalBanachspaces(Theorem5.1)
Denjoy-Afterthepreliminarysubmission,thankstotheReferee,wehaveknownabouttheexistenceof 4,12,14],whicharequiterelatedtothispaper.In 4] theauthorscontinuedtheirresearchin 3] onBergmanspacesA p
α,weightedBanachspacesH v α withstandardweights,theclassicalDirichletspaceD,andBlochspacesB v α.Some resultsconcerning ergodic propertiesof composition operators onthe Hardy spaceH p were proved
in 12,14].Moreover, in 14] theauthors alsostudied(uniformly)meanergodicityofC ϕon(non-weighted)Bergmanspaces A p,Bloch spacesB v α, weightedBanachspaceswith sup-normH v and H0 Obviously, allresultsin 4,12,14] andthis paperhavebeen obtainedindependently withdifferent approaches Indetails,themethodin 4,14] isbasedontheessentialspectralradiusr e (C ϕ) and 3,Theorem3.4],whileweusetheCarlesonmeasureforBergmanspacesA p α,DirichletspacesD p
αandthecharacteristicsofweightsforweightedBanachspaceswithsup-norm,weightedBlochspaces.ItshouldbenotedthatallresultsonDirichletspaces
D p
αandweightedBlochspacesB v inthispaperarenewandnotcoveredbytheother papers;furthermore,Theorem4.8andExample4.9improveandcomplementtheresultsin 14] forweightedBanachspaceswithsup-norm(seeRemark5.2 formoredetails)
Constants Throughout this paper we use thenotation A B for nonnegative quantities A and B to
mean that there is an inessential constant C > 0 such that A ≤ CB. Similarly, the notation A
meansthatboth A B and B A hold. Moreover,forsimplicity, wewrite C z0 insteadofC ϕ inthecase
ϕ(z) ≡ z0∈ D.
2 Preliminaries
InthissectionwerecallthedefinitionoftheBanachspacesofholomorphicfunctionswhichwillbestudiedbelowandincludesomeresultsoncompositionoperators inthem
LetX ⊂ H(D) be aBanachspaceendowedwith itspropernorm · X.If theinclusioni : X → H(D)
iscontinuous,then X issaidtobe aBanach space of holomorphic functions andwewrite X → H(D) 2.1 Hardy spaces H p , Bergman spaces A p
wheredθ is theLebesguemeasure on[0, 2π].
Next,givenα > −1 and1≤ p < ∞,theBergmanspaceA p α isdefinedasfollows:
Trang 19The Dirichletspace D p
α consists of allholomorphicfunctions onD whosederivativesbelong toA p
α.Notethattheset P ofallpolynomialsiscontainedanddense
inthesespacesH p,A p α andD p
α.Foreacha ∈ D,letφ adenote thedisk automorphisminterchanging a and0,i.e
α-Carleson measure,iftheembeddingoperatorI μ : A p
α → L p (D, dμ) isbounded,i.e.thereexistsaconstant
Inthiscasewewriteμ α=I μ p
A p α →L p (D,dμ).Accordingto 24,Theorems2.25and 2.26],apositiveBorelmeasure μ isanα-Carlesonmeasure ifandonlyif,forsomenumberr ∈ (0,1),
by
(ωA α)◦ ϕ −1 (E) :=
ϕ −1 (E)
ω(ζ)dA α (ζ)
for everyBorelset E ⊂ D.Then, by thechange-of-variableformula of measuretheory (see, 11,Theorem
C,p 163]), foreachpositive Borelfunctionh on D,weget
Trang 20|f (z) | 1
(1− |z|2)α+2 p +1f A p α (2.5)
By 24,Lemmas1.24and2.20],
1− |z|2 2 for all ζ ∈ Δ(z) and z ∈ D. (2.6)
2.2 Weighted Banach spaces H v , weighted Bloch spaces B v
Recallthataradial weight (briefly,aweight)onD isapositivefunctionv onD withv(z) = v( |z|), z ∈ D,
wherev(r) iscontinuousand increasingon[0,1) andv(r) → ∞ as r → 1 −.Foraweightv onD,wedefine
thefollowingweightedBanachspaceof holomorphicfunctionsonD:
Thisspaceisalso knownasaweightedBergman spaceofinfiniteorder
TheweightedBloch spaceB v consistsof allholomorphicfunctionswhose derivativesbelong toH v and
Trang 21It iswell known(see, forinstance, 1,2])thatto characterize topologicalpropertiesof thespaces H v and
H0orlinearoperatorsbetweenthemintermofweights,thestandardwayistousetheso-calledassociatedweights.Accordingto 6,Definition 1.1],foraweightv onD,itsassociated weight isdefinedby
v(z) := sup{|f(z)| : f ∈ H v (D), f v ≤ 1}.
Notethatv(z)=v(|z|) and0< v(z) ≤ v(z) forallz ∈ D, v(r) isincreasingandlog-convex on [0,1) (i.e.thefunctionlogv(e x) is convexon(−∞,0)), andH v(D)= H v(D) isometrically.Moreover, in 13, Lemma2.2]
it wasshownthateverylog-convex weightv onD isessential,thatis, thereissomeconstantM suchthat
v(z) ≤ v(z) ≤ Mv(z), z ∈ D. Thus,forourmainpurposesitseemsreasonable touselog-convexweights.Moreover, for alog-convex weight v,by 8, Theorem 2.3] the following condition from 16, p 310] and[17,Definition2.1]
lim sup
k →∞
v(1 − 2 −k−1)v(1 − 2 −k) < ∞ (2.7)
is equivalenttothecontinuityofallcompositionsoperators C ϕ , ϕ ∈ S(D),onH v andH0.Note thatsomeconditionsofvarious typesthatareequivalentto (2.7) werestatedin 1,Lemma2.6]
TheboundednessofC ϕ onweightedBlochspacesB v ischaracterized asfollows
Proposition 2.1.Let v be a log-convex weight on D and ϕ ∈ S(D) Then the following statements are valid.
(a) The operator C ϕ is bounded on B v if and only if
sup
z ∈D
v(ϕ(z)) |ϕ (z) | v(z) < ∞.
(b) Each automorphism φ a , a ∈ D, induces a bounded operator C φ a on B v if and only if v satisfies condition
(2.7).
Proof The part(a)follows from 19,Theorem 2.3]
(b) First, suppose thatv satisfies condition (2.7) Then by 8, Corollary 2.2 and Theorem 2.3], C φ a isbounded onH v and sup
1− |a|supz∈D
v(φ a (z))
v(z) < ∞.
Thus, bypart(a),C φ a isboundedonB v
Next,suppose thatforeverya ∈ D,theoperatorC φ a isboundedonB v.Then,bypart(a),weget
∞ > sup
z ∈D
v(φ a (z)) |φ
a (z) | v(z) ≥1− |a|
1 +|a|supz ∈D
v(φ a (z))
v(z) .
From thisandtheproofof 8,Theorem 2.3],itfollowsthatv satisfiescondition(2.7)
Inviewoftheabovereasons,throughoutthispaperweonlyconsidercompositionoperatorsonthespaces
H v and B v given by log-convex weights v satisfying (2.7) Let V denote the set of all such weights Thestandardweights v α (z):= (1− |z|2)−α , α > 0,belongtoV.
Trang 223 Elliptic automorphismsymbol
Inthis section westudy the sequenceC ϕ n inducedby anelliptic automorphism ϕ ∈ S(D).Recall that
ϕ ∈ S(D) isanellipticautomorphismifandonlyifthereexistλ ∈ ∂D and a ∈ D suchthatϕ = φ a ◦ ψ ◦ φ a,whereψ(z) = λz, z ∈ D;moreover,ϕ isoforderk ifandonlyifλ isthek-throotofunityforsomesmallestnumberk ∈ N, andϕ isofinfiniteorder ifandonlyifλ isnotarootof unity
The main result in this case is stated in the following theorem, which should be compared with 14,Proposition 2.5]
Theorem 3.1.Let X be a Banach space of holomorphic functions such that the set P of all polynomials is dense in X, all operators C φ a with a ∈ D are bounded on X, and all operators C λz with λ ∈ ∂D are power bounded on X Suppose that ϕ is an elliptic automorphism of D Then the sequence C ϕ n does not converge
in the WOT Furthermore,
(a) if ϕ is of order k, then C ϕ is uniformly mean ergodic on X.
(b) if ϕ is of infinite order, then C ϕ is mean ergodic on X, but not uniformly mean ergodic on X.
Proof Sinceϕ = φ a ◦ ψ ◦ φ aforsomea ∈ D and ψ(z) = λz with λ ∈ ∂D, C ϕ n = C φ a ◦ C ψ n ◦ C φ a forevery
n ∈ N. Fromthisandthehypothesis,itissufficientto considerthecasewhenϕ(z) = λz with λ ∈ ∂D.
(a) Supposethatλ k = 1 with somesmallest numberk ∈ N. Inthis case thesequence C ϕ n is periodic,andhence,itdoesnotconvergeintheWOT
Foreachf ∈ X and n ∈ N,writingn = mk + s with0≤ s < k,putting
Consequently,C ϕisuniformly meanergodiconX.
(b) Suppose now λ is nota root of unity In this case, for the evaluation δ z0 at z0 D and thefunctione1(z) := z,wegetthatthesequenceδ z0(C ϕ n e1)= λ n z0 cannotbeconvergent.Thus,C ϕ ndoesnotconvergeintheWOT
Trang 23Next,foreachfunctione m (z) := z m withm ∈ N,we have
Moreover, for the function e0(z) := z0, (C ϕ)[n] (e0) = C0(e0) = 1 for every n ∈ N. Thus, the sequence
(C ϕ)[n]convergesto C0 ontheset ofallmonomials, andhence,onthesetofallpolynomials.Ontheotherhand,sinceC ϕ ispowerboundedonX,thesequence (C ϕ)[n] isequicontinuousonX.Consequently,C ϕ ismean ergodiconX by 5,Lemma 2.1] andthedensityoftheset ofallpolynomialsinX.
Now we will prove by contradiction that C ϕ is not uniformly mean ergodic on X. Indeed, if C ϕ wasuniformly meanergodiconX,then
Im(I − C ϕ ) = KerC0= X0,
whereX0:={f ∈ X : f(0)= 0} isaclosedsubspaceofX.Thismeansthatforeachf (z)=∞
j=1 a j z j ∈ X0,there exists a functiong(z)=∞
j=0 b j z j ∈ X such thatg − g ◦ ϕ = f Thatis, g(z) − g(λz) = f (z), andhence,
Trang 24Ontheotherhand,X0isaBanachspaceendowedwiththenorm · X.Fromthisandtheclosedgraphtheorem,theoperatorT isboundedonX0, thatis,T f X ≤ T L(X0 )f X for allf ∈ X0.Inparticular,foreachm ∈ N,
whichisacontradiction,sincethesequence(λ m)misuniformlydistributedon∂D.
Remark3.2 Theresults inTheorem 3.1 hold onHardy spaces H p, Bergman spaces A p α, Dirichletspaces
D p
α, weightedBanachspacesH0 andweightedBloch spacesB0 foreveryp ≥ 1, α > −1 andeveryweight
v ∈ V.
Concerning weightedBanachspacesH0 and H v, in 5, Theorem A.1] the authors constructedaweight
v α so thatthe operatorC ϕ induced byϕ(z) = λz,where λ is notaroot of unity, is notuniformly meanergodic onH0
α, andhence, notmean ergodic onH v α.From Theorem 3.1 it followsthat theoperatorC ϕ
inducedbyanellipticautomorphismϕ ofinfiniteorderisnotuniformlymeanergodiconH0andnotmeanergodiconH v foreveryweightv ∈ V.Inparticular,ifϕ(z) = λz,whereλ ∈ ∂D isnotarootofunity,then,
bytheproofof Theorem3.1,theoperatorC ϕ isnotuniformlymeanergodic onH0 andnotmeanergodic
onH v foreveryweightv onD
4 Symbolwithinterior Denjoy-Wolffpoint
In this section we show thatif the Denjoy-Wolff point of ϕ is in D, then the sequence C ϕ n convergesuniformlyformanyspacesX.However,thetechniqueonvariousspacesareessentiallydifferentinthiscase,
sowe willseparatethestudyforeachspace
NotethatiftheDenjoy-Wolffpointofϕ is z0 inD,thenthefunctionψ = φ z0◦ ϕ ◦ φ z0 fixes0 inD and
C ϕ n = C φ z0 ◦ C ψ n ◦ C φ z0 for every n ∈ N. This and thefact thatall operators C φ a are bounded on theconsideredspacesX allowustosuppose thattheDenjoy-Wolffpointofϕ is0 throughoutthissection
Wewillusethefollowinglemma inthesequel.Foreachϕ ∈ S(D), r ∈ (0,1),andn ∈ N,weput
Proof By 10, Lemma 7.33], there is anumber A > 1 such that1− |ϕ(z)| > A(1 − |z|) for every z ∈ D
with|z| ≥ r.Then, foreveryz ∈ E c
n,r,bytheSchwarzlemma,we get
Trang 254.1 On Hardy spaces H p
In 3,Theorems 4.8,4.10and4.12] theauthorsprovedthatforeveryp ≥ 1 and ϕ ∈ S(D) with Wolffpointz0inD,thesequenceC ϕ n alwaysconvergesintheWOT,andhence,C ϕisalwaysmeanergodic
Denjoy-on H p; moreover C ϕ n converges uniformly if ϕ isnot inner, and it doesnot converge in the SOTif ϕ is
inner.Inviewofthis,weareonlyinterestedintheuniformmeanergodicityoftheoperatorC ϕassociated
onH0p.Thus,I − C ϕ isaninvertibleoperatorfrom H0p intoitself.This meansthat1∈ σ(C / ϕ | H0p)
Ontheotherhand,by 10,Theorem3.8],C ϕ isanisometryonH p,andhence,onH0p.Fromthisand 9,Exercise 7, p.213],it followsthateitherσ(C ϕ | H p0)⊂ ∂D if C ϕ | H0p isinvertible or σ(C ϕ | H0p)= D ifC ϕ | H0p
Trang 26Ontheotherhand,using(2.1) and(2.4),weobtain
|f(ϕ n (z)) | p dA α (z) +
E c n,r
dA α (z)
f p
A p α (χ E c n,r A α)◦ ϕ −1
n α+f p
A p α A α (E n,r c ),
whereχ E isthecharacteristicfunctionofasubsetE in D
Consequently,foreachn ∈ N,
n,r)→ 0 as n → ∞.Moreover,fromthisandtheLebesgue’sdominatedconvergencetheorem,
italsofollowsthat
Itremainstoprovethatthelastlimitis0.Todothis,wetakeanumberγ ∈ (−1, α) andputβ := α −γ > 0.
Foreveryz ∈ D, using(2.6),weget
ϕ −1 n (Δ(z))
χ E c n,r (ζ) (1− |ζ|2)β
(1− |ϕ n (ζ) |2)β(1− |ϕ n (ζ) |2)β dA γ (ζ)
≤
sup
ζ ∈D χ E
c n,r (ζ) (1− |ζ|2)β(1− |ϕ n (ζ) |2)β
ϕ −1 n (Δ(z))
(1− |ϕ n (ζ) |2)β dA γ (ζ)
sup
ζ ∈D χ E
c n,r (ζ) 1− |ζ|2
ζ ∈E c n,r
Trang 27(χ E c n,r A α)◦ ϕ −1
n α= sup
z ∈D
(χ E c n,r A α)◦ ϕ −1
n (Δ(z))
(1− |z|2)α+2
sup
ζ∈E c n,r
ζ∈E c n,r
ζ ∈E c n,r
ζ∈E c n,r
1− |ζ|2
1− |ϕ n (ζ) |2
β
→ 0 as n → ∞,
where thelastlimitisbasedonLemma4.1
Theproofis completed
From Theorem 4.3 and 3, Theorem 3.4], we immediately get the following result, which extends thecorresponding onein 18, Section5] to allBergmanspaces A p α andfunctionsϕ ∈ S(D) withDenjoy-Wolffpoint inD
Corollary 4.4.Let α > −1, p ≥ 1, and ϕ ∈ S(D) with Denjoy-Wolff point in D Then the essential spectral radius r e,A p
whichmeansthatD p
αandA p α −parehomeomorphic.FromthisandTheorem4.3,itfollowsthatifp ≥ 1 and
α > p − 1,thenforeveryfunctionϕ ∈ S(D) withDenjoy-Wolffpoint0,thesequenceC ϕ n convergesto C0
Proof We dividetheproofinto severalsteps
Step 1 Weshow thatC ϕ is boundedon D p
α Put β := α − γ > 0.For everyz ∈ D, using theSchwarzlemma and(2.6),weget
Trang 28α.Indeed,foreveryr ∈ (0,1) andn ∈ N, weget
Trang 29Step3 WeshowthatC ϕ n − C0 L(D p
α)→ 0 as n → ∞.Fixanumberr ∈ (0,1).For everyf ∈ D p
n | p A α)◦ ϕ −1
n α f p
A p α
≤ (χ E c n,r |ϕ
n (ζ) | p dA α (ζ)
ϕ −1 n (Δ(z))
χ E c n,r (ζ) (1− |ζ|2)β(1− |ϕ n (ζ) |2)β(1− |ϕ n (ζ) |2)β |ϕ
n (ζ) | p dA γ (ζ)
≤
sup
ζ∈D χ E
c n,r (ζ) (1− |ζ|2)β
ζ∈D χ E
c n,r (ζ) 1− |ζ|2
Trang 30This,Lemma4.1and(2.3) implythat
ζ ∈E c n,r
ζ∈E c n,r
ζ∈E c n,r
1− |ζ|2
1− |ϕ n (ζ) |2
β
C ϕ n p L(D p
)
≤
sup
n C ϕ n p
L(D p
)
sup
ζ ∈E c n,r
1− |ζ|2
1− |ϕ n (ζ) |2
β
→ 0 as n → ∞,
sinceC ϕispowerboundedonD p
Theproofiscompleted
Note thatit isnoteasyto check thatthe operatorC ϕ ispowerbounded onDirichletspaces D p.Sowewillconsider thefollowingparticularcases,inwhichthisassumptionisfulfilled
Corollary4.6 Let p, α satisfy (4.1) and ϕ ∈ S(D) with Denjoy-Wolff 0 such that supn ϕ
Proof In this case we will show thatC ϕ is power bounded onD p
α for every α > −1. Then the assertionfollowsfrom Theorem4.5
Foreveryn ∈ N and f ∈ D p
α,byLittlewood’sSubordinationTheorem,weget
Trang 31Proof Letγ < α sothat2≤ p ≤ γ + 2.ByTheorem4.5,itissufficienttoprovethatC ϕispowerbounded
onD p.Foreveryn ∈ N and f ∈ D p, usingtheSchwarzlemmaandPick-Schwarzlemma,weget
4.4 On weighted Banach spaces H v
In 22, Theorems 10 and 11] itwas shownthatthe operatorC ϕ inducedby afunction ϕ ∈ S(D) with
Denjoy-WolffpointinD isuniformlymeanergodiconH v.However,theproofwasstatedformodelfunctions
ϕ(z) = λz with |λ| < 1 or ϕ(z) = z n withn ≥ 2.Inthissubsectionweimprovethisresultanditsproofs.Theorem 4.8.Let v be a weight in V and ϕ ∈ S(D) with Denjoy-Wolff point 0 The following statements are true:
(a) The sequence C ϕ n converges in the SOT to C0 on H0.
then the sequence C ϕ n converges to C0 in L(H v ).
Proof First, weshowthatϕ n → 0 in H v as n → ∞.Indeed,foreveryr ∈ (0,1) andn ≥ N,wehave
ϕ n H v = sup
z∈D
|ϕ n (z) | v(z) = max
#sup
|z|≤r
|ϕ n (z) | v(z) , sup |z|>r
|ϕ n (z) | v(z)
$
≤ max
#1
Lettingn → ∞ and thenr → 1 − inthelast inequality,wecangetϕ n H v → 0 as n → ∞.
(a) Fixanarbitraryfunctionf ∈ H0.Foreveryr ∈ (0,1) andn ∈ N,weobtain
C ϕ n f − C0f H v = max
#sup
Trang 32From 1,Theorem 2.8] itfollowsthat
Thus,theassertionfollows
(b)Fixanarbitrarynumberr ∈ (0,1).Similarlytopart(a),foreveryn ∈ N and f ∈ H v,wehave
C ϕ n f − C0f H v = max
#sup
(1− |ϕ n (z) |) α v(ϕ n (z)) (1 − |z|) α v(z)
foreveryn ∈ N and z ∈ D.Thus,
Trang 33z ∈E c n,r
|f(ϕ n (z)) − f(0)|
v(z) ≤ f H v sup
z ∈E c n,r
sup
z∈E c n,r
sup
z∈E c n,r
1− |z|2
1− |ϕ n (z) |2
α
Consequently,foreveryr ∈ (0,1) andn ∈ N,
z ∈E c n,r
Theproofis completed
Thefollowingexampleshowsthatpart(a)inTheorem4.8doesnotholdforthespaceH vandcondition(4.3) isessentialforpart(b)inthistheorem
Example 4.9.Letλ ∈ (0,1) andconsider
ϕ(z) := λz
1− (1 − λ)z and v(r) := log
e
1− r .
Obviously, ϕ ∈ S(D) with Denjoy-Wolff 0, the function f0(z) := log 1
1− z belongs to H v, but does notbelongto H0,and
lim sup
r→1 −
(1− r)v (r) v(r) = 0,
which, by 1, Lemma2.6], impliesthatv ∈ V, and hence,the operatorC ϕ is bounded onH v,but v does
notsatisfycondition(4.3).Moreover,foreachn ∈ N,
log1−(1−λ n )z
1−z log1−|z| e
≥ sup
r ∈(0,1)
log1−(1−λ1−r n )rlog1−r e ≥ sup
Trang 344.5 On weighted Bloch spaces B v
Theorem4.10.Let v be a weight in V satisfying
lim inf
r →1 −
(1− r)v (r)
Then for every ϕ ∈ S(D) with Denjoy-Wolff point 0, the sequence C ϕ n converges to C0 in L(B v ).
Proof Wedividetheproofinseveral steps
Step1 WeprovethattheoperatorC ϕisboundedonB v.By(4.4),therearenumbersγ > 1 and r0∈ (0,1)suchthat
(1− r) γ v (r) − γ(1 − r) γ −1 v(r) > 0
forallr ∈ [r0,1).Hence,thefunction(1− r) γ v(r) isincreasingon[r0,1),whichimpliesthat
(1− r1)γ v(r1) (1 − r2)γ v(r2)foreveryr1≤ r2 in[0,1).Usingthis, theSchwarzlemma andPick-Schwarzlemma,weget
sup
z ∈D
v(ϕ(z)) |ϕ (z) | v(z) ≤ sup
Thus,theassertionfollowsfrom Proposition2.1
Step2 Weshow thatϕ n converges to0 inB v Indeed,foreachn ∈ N and r > r0,bythePick-Schwarzlemma,
= max
#sup
|z|≤r
|ϕ
n (z) | v(z) , sup |z|>r
|ϕ
n (z) | v(z)
$
≤ max
#sup
|z|≤r
|ϕ
n (z) | v(z) , sup |z|>r
v(0) |z|≤rsup|ϕ
n (z) |, sup
|z|>r
1(1− |z|)v(z)
$
≤ max
#1
Lettingn → ∞ andthenr → 1 − inthelastinequality,wegetthatϕ n B v → 0 as n → ∞.
Step3 WeprovethatC ϕ n − C0 L(B v)→ 0 as n → ∞.Foreveryn ∈ N and f ∈ B v, wehave
C ϕ n f − C0f B v = sup
z∈D
|f (ϕ
n (z))ϕ n (z) | v(z) = max
#sup
z ∈E
|f (ϕ
n (z))ϕ n (z) | v(z) , sup z ∈E c
|f (ϕ
n (z))ϕ n (z) | v(z)
$
,
Trang 35z∈E n,r0
v(ϕ n (z)) |ϕ
n (z) | v(z) ≤ f B v v(r0)ϕ n B v ,
and, bythePick-Schwarzlemmaagain,
z∈E c n,r0
v(ϕ n (z)) |ϕ
n (z) | v(z) ≤ f B v sup
z∈E c n,r0
v(ϕ n (z))(1 − |ϕ n (z) |2)
v(z)(1 − |z|2) .Consequently,foreveryn ∈ N,
C ϕ n − C0 L(B v)≤ max
#
v(r0)ϕ n B v , sup
z∈E c n,r0
v(ϕ n (z))(1 − |ϕ n (z) |2)
v(z)(1 − |z|2) .
It remainstocheck thelast limitis0.Indeed,foreveryz ∈ E c
n,r0,bytheSchwarzlemma,
v(ϕ n (z))(1 − |ϕ n (z) |) v(z)(1 − |z|) ≤ sup z∈E c
z ∈E c n,r0
1− |z|2
1− |ϕ n (z) |2
γ−1
→ 0 as n → ∞.
Thefollowing exampleshowsthatcondition(4.4) inTheorem4.10isessential
Example 4.11.Letϕ(z) = z2 andv(r)= 1
1− r.Obviously, v doesnotsatisfycondition(4.4) and
sup
z∈D
v(ϕ(z)) |ϕ (z) | v(z) = supz∈D
|2z|(1 − |z|)
1− |z|2 = 1.
Thus, C ϕ isboundedonB v byProposition2.1
Obviously, C ϕ n z → C0z in H(D) as n → ∞.However,foreachn ∈ N,
C ϕ n z − C0z B v = sup
z ∈D
|ϕ
n (z) | v(z) = 2
Trang 365 SymbolwithboundaryDenjoy-Wolffpoint
In 3,Theorem 4.3] itwasshownthattheoperatorC ϕinducedbyafunctionϕ ∈ S(D) withDenjoy-Wolffpointz0∈ ∂D isnotpowerboundedonsomespacesX,suchasHardyspacesH p,spaceszH pandweightedHardyspacesH p (β) with p ≥ 1.InthissectionweextendthisresulttootherBanachspacesofholomorhicfunctions.Moreover,wealsoprovethatinthiscasetheoperatorC ϕisnotmeanergodiconthesespaces.Theorem 5.1 Let X be a Banach space of holomorphic functions and ϕ ∈ S(D) with Denjoy-Wolff point
z0∈ ∂D.
(a) If there is a function f0 ∈ X such that |f0(z) | → ∞ as z → z0 in D, then C ϕ is not power bounded
on X.
(b) If the functionlogz01−z belongs to X, then C ϕ is not mean ergodic on X.
In particular, the following statements are valid:
(i) For every p ≥ 1, C ϕ is not power bounded and not mean ergodic on H p
(ii) For every p ≥ 1 and α > −1, C ϕ is not power bounded and not mean ergodic on A p
α
(iii) For every1≤ p < α + 1, C ϕ is not power bounded and not mean ergodic on D p
α
(iv) For every weight v ∈ V satisfying log1−r1 = O(v(r)) as r → 1 − , C
ϕ is not power bounded and not mean ergodic on H v
(v) For every weight v satisfying 1−r1 = O(v(r)) as r → 1 − , if C
ϕ is bounded on B v , then C ϕ is not power bounded and not mean ergodic on B v
Proof (a) SinceX → H(D),
|f(0)| f X for all f ∈ X. (5.1)
In particular, |f0(ϕ n(0))| C ϕ n f0 X for everyn ∈ N.Since ϕ n(0) → z0 as n → ∞,by thehypothesis,
|f0(ϕ n(0))| → ∞ as n → ∞. Thus,C ϕ isnotpowerbounded
sinceϕ n(0)→ z0 asn → ∞. Thisimpliesthatthesequence ((C ϕ)[n] h) n doesnotconverge inX.Thus,C ϕ
isnotmeanergodiconX.
Toshow (i)–(v),it issufficientto provethatthefunctionh(z)= logz1
0−z belongsto thecorresponding
2 sinθ2
p
dθ < ∞.
Trang 37(ii)Putting z0:= e iθ0 andbythedefinitionof · A p α,weget
1− r
p
(1− r2)α rdr < ∞.
(iii) Theassertionfollowsfrom (ii)andthefactthatD p
αandA p α −p arehomeomorphicwhenα > p − 1.
(iv)Bythedefinitionof · H v andlog1−r1 = O(v(r)) as r → 1 −,itimmediatelyimpliesthath ∈ H v.(v)Since 1−r1 = O(v(r)) as r → 1 −,h (z) ∈ H v.Then,h ∈ B v
Remark 5.2 As mentioned in the Introduction, someresults in this paper for composition operators C ϕ
on Hardyspaces H p, Bergman spacesA p
α, andweightedBanachspaces with sup-normhavealready beenestablishedindependentlyin 4,12,14].Indetails:
• OnHardyspacesH p:in 14,Proposition 2.5andTheorem 3.12] and 12,Theorem 8],theauthors alsoshowedtheresultsinTheorem3.1(forHardyspacesH p)and Theorem4.2bydifferentapproaches
• OnBergman spacesA p α:in 4, Theorem 2.9],Arendt,Chalendar,Kumarand Srivastava firstlyprovedthatthe essential spectral radius r e,A p
α (C ϕ) isstrictly lessthan 1, and then used 3, Theorem 3.4] toconcludethatthesequenceC ϕ n convergesuniformlyonA p
Further-holomorphic functions, but these conditions are not necessary for the proof of the parts related toTheorem5.1
However, itisworth emphasizingthatall resultsonDirichletspaces D p
α and weightedBloch spacesB v
inthis paper, andcomposition operatorsC ϕ inducedby anellipticautomorphism onBergman spacesA p
α
havenotyetbeenconsideredintheotherpapers
Moreover, Theorem4.8 essentiallyimprovestheresultsobtainedin 14] for weightedBanachspacesH v
and H0 Infact,Jordáand Rodríguez-Arenasprovedthatifv isaconvenient weight onD (i.e.v satisfies
(2.7) andthereisf ∈ H0suchthatRef (z) → +∞ as z → 1)andϕ ∈ S(D) hasaDenjoy-WolffpointinD,thenthesequenceC ϕ n alwaysconvergesintheWOTonH v0([14,Theorem 3.7])andC ϕisuniformlymeanergodic onH v0 and/orH v ifandonlyifC ϕ isquasicompact([14,Theorem 3.8]).Whileweprovedthatforeveryweightv ∈ V andeveryϕ ∈ S(D) with Denjoy-Wolffpoint inD,thesequence C ϕ n alwaysconverges
inthe SOTonH0 andconverges uniformly onH v undersomenaturalcondition(4.3) (see Theorem 4.8)
On theotherhand,itis notdifficult tocheck thatthefunctionϕ andtheweightv inExample 4.9satisfyallconditionsin 14,Proposition 3.14].Then,by 14,Theorem 3.8],thisoperatorC ϕisnotuniformlymeanergodic on H0, and hence,on H v Therefore, C ϕ is not mean ergodic on H v This and the fact that C ϕ
is power bounded on H v also imply thatthe sequence C ϕ n does not converge even inthe WOT on H v;however,byTheorem4.8(a),thissequenceconverges intheSOTonH0
Trang 38Apartofthispaperhasbeendoneduringtheauthor’sstayattheVietnamInstituteforAdvancedStudy
inMathematics.Hewouldlike tothanktheinstitutionforhospitalityandsupport
The author is deeply grateful to the Referee for alerting about the existence of 4,12,14], the usefulremarks andcommentsthatled tothe improvementof thispaper Theauthor is alsogreatly indebtedbyProf.E Jordáforsharinghis paper 14]
[5] M.J Beltrán-Meneu, M.C Gómez-Collado, E Jordá, D Jornet, Mean ergodic composition operators on Banach spaces
of holomorphic functions, J Funct Anal 270 (2016) 4369–4385.
[6] K.D Bierstedt, J Bonet, J Taskinen, Associated weights and spaces of holomorphic functions, Stud Math 127 (1998) 137–168.
[7] J Bonet, P Domański, A note on mean ergodic composition operators on spaces of holomorphic functions, Rev R Acad Cienc Exactas Fís Nat., Ser A Mat 105 (2011) 389–398.
[8] J Bonet, P Domański, M Lindström, J Taskinen, Composition operators between weighted Banach spaces of analytic functions, J Aust Math Soc A 64 (1998) 101–118.
[9] J.B Conway, A Course in Functional Analysis, Springer-Verlag, New York, 1985.
[10] C.C Cowen, B.D MacCluer, Composition Operators on Spaces of Analytic Functions, Studies in Advanced Mathematics, CRC Press, Boca Raton, FL, 1995.
[11] P Halmos, Measure Theory, Springer-Verlag, New York, 1974.
[12] S-A Han, Z-H Zhou, Mean ergodicity of composition operators on Hardy space, Proc Indian Acad Sci Math Sci 129 (4) (2019), Art 45, 10 pp.
[13] O Hyvärinen, M Kemppainen, M Lindström, A Rautio, E Saukko, The essential norm of weighted composition operators
on weighted Banach spaces of analytic functions, Integral Equ Oper Theory 72 (2012) 151–157.
[14] E Jordá, A Rodríguez-Arenas, Ergodic properties of composition operators on Banach spaces of analytic functions, J Math Anal Appl 486 (2020) 123891.
[15] M Lin, On the uniform ergodic theorem, Proc Am Math Soc 43 (1974) 337–340.
[16] W Lusky, On weighted spaces of harmonic and holomorphic functions, J Lond Math Soc 51 (1995) 309–320.
[17] W Lusky, Growth conditions for harmonic and holomorphic functions, in: S Dierolf, et al (Eds.), Functional Analysis,
[20] K Petersen, Ergodic Theory, Cambridge University Press, Cambridge, 1983.
[21] J.H Shapiro, Compositions Operators and Classical Function Theory, Springer-Verlag, New York, 1993.
[22] E Wolf, Power bounded composition operators, Comput Methods Funct Theory 12 (2012) 105–117.
[23] K Yosida, Functional Analysis, Springer, Heidelberg, 1965.
[24] K Zhu, Spaces of Holomorphic Functions in the Unit Ball, Springer-Verlag, New York, 2005.
Trang 39Contents lists available at ScienceDirect
Journal of Functional Analysis
aVNU University of Science, Vietnam National University, Hanoi, Vietnam
bGraduate School of Mathematical Sciences, The University of Tokyo, 3-8-1
Komaba, Meguro-ku, Tokyo 153-8914, Japan
This work is dedicated to Professor
Nguy˜ ên Viê.t D˜ ung on the occasion
Higher order fractional Laplacian
Gaussian hypergeometric function
Inspired by a recent sharp Sobolev trace inequality of order four on the balls Bn+1found by Ache and Chang (2017) [ 2 ], we propose a different approach to reprove Ache–Chang’s trace inequality To further illustrate this approach, we reprove the classical Sobolev trace inequality of order two on Bn+1 and provide sharp Sobolev trace inequalities of orders six and eight
on Bn+1 To obtain all these inequalities up to order eight, and possibly more, we first establish higher order sharp Sobolev trace inequalities onRn+1+ , then directly transferring them to the ball via a conformal change As the limiting case of the Sobolev trace inequalities, Lebedev–Milin type inequalities of order up to eight are also considered.
© 2019 Elsevier Inc All rights reserved.
* Corresponding authors.
E-mail addresses:nqanh@vnu.edu.vn , ngo@ms.u-tokyo.ac.jp (Q.A Ngô), vanhoang0610@yahoo.com ,
nvhoang@math.ac.vn (V.H Nguyen), phanquochung@dtu.edu.vn (Q.H Phan).
1 Funded by the Vietnam National University, Hanoi under project number QG.19.12 and by the Tosio Kato Fellowship awarded in 2018.
2 Partially funded by CIMI’s postdoctoral research fellowship and by the Simons Foundation Grant Targeted for Institute of Mathematics, Vienam Academy of Science and Technology.
https://doi.org/10.1016/j.jfa.2019.108414
0022-1236/© 2019 Elsevier Inc All rights reserved.
Trang 40Sobolev trace inequality, Beckner
inequality, Lebedev–Milin inequality
4 Sobolev trace inequality of order four: proof of Theorem 1.2 26 4.1 Sharp Sobolev trace inequality of order four on Bn+1: proof of Theorem 1.2 26 4.2 A Beckner type inequality of order four on Bn+1 31 4.3 A Lebedev–Milin type inequality of order four on B 4 34
5 Sobolev trace inequality of order six 35 5.1 Neumann boundary condition for extensions 36 5.2 Sharp Beckner type inequality of order six on Bn+1 39 5.3 Sharp Sobolev trace inequality of order six on Bn+1: proof of Theorem 1.3 42 5.4 A Lebedev–Milin type inequality of order six on B 6 43
6 Sobolev trace inequality of order eight and beyond 44 Acknowledgments 47 Appendix A Proof of Proposition 3.1 48 References 51
Of importance in analysis and conformal geometry are Sobolev and Sobolev traceinequalities either on Euclidean spaces or on Euclidean balls These inequalities, inbrief,providecompactembeddingsbetweenimportantfunctionalspaces.FortheclassicalSobolevinequality(ofordertwo),itsversiononRn isgivenasfollows
Γ(n+22 )Γ(n−22 )ω
n−2 n
Rn
|∇u|2dz (1.1)
for anysmooth functionu withcompactsupport.Here,andthroughout thispaper, ω n
isthevolumeoftheunitsphereSn,theboundaryoftheunitballBn+1,inRn+1,which
is 2π (n+1)/2 /Γ((n + 1)/2), which is also 2n π n/2 Γ(n/2)/Γ(n). It is well-known that theinequality(1.1) iscrucial intheresolution oftheYamabeproblemonclosedmanifolds.Not limitedto theYamabeproblem,Inequality(1.1) is thefundamentaltool and have