You are currently browsing the category archive for the '[18++]Complex Analysis' category.

Không nghi ngờ gì rằng mặt Riemann là một trong các  đối tượng quan trong nhất của Toán học, ta thấy là nó xuất hiện khắp nơi, trong Giải tích, Hình học và Lý thuyết số, nó sinh ra khi người ta xét các bài toán của Giải tích phức mà ở đó sinh ra hàm đa trị, kiểu như là một giá trị của biến lại có nhiều hơn một giá trị của hàm. Trong bài này tôi sẽ giới thiệu một vài ví dụ để làm rõ điều này.

1, Các hàm đại số

Trong giải tích phức, hàm đa trị đầu tiên ta gặp phải là hàm f(z)=\sqrt{z} (mỗi số phức khác 0 có hai căn bậc hai khác nhau), đây chính là một trường hợp đặc biệt của hàm đại số , nghĩa là hàm f mà các giá trị của nó là nghiệm của một đa thức có hệ số phân hình. Ta có kết quả tổng quát sau:

Gỉa sử X là một mặt Riemann và P(T)=T^n+c_1T^{n-1}+\cdots+c_n\in\mathcal{M}(X)[T] là một đa thức bất khả quy có bậc n. Khi đó có mặt Riemann Y, một ánh xạ phủ chỉnh hình rẽ nhánh n-\pi :Y\to X và một hàm phân hình F\in\mathcal{M}(Y) sao cho (\pi^*P)(F)=0. Bộ ba (Y,\pi,F) là duy nhất theo nghĩa tự nhiên.

Trong định lý trên có rất nhiều khái niệm khó, bạn đọc muốn tìm hiểu chính xác thì hãy xem GTM81.  Ở trên ta thấy rõ là mới đầu có một đa thức với hệ số phân hình bậc n, người ta muốn tìm một hàm là nghiệm của phương trình này, vì mỗi giá trị của biến sẽ có n giá trị của T theo định lý cơ bản của Đại số, nên vấn đề đa trị nảy sinh. Đó là lúc mặt Riemann Y xuất hiện, và nó giải quyết chọn vẹn vấn đề này. Hình thức của Y như là nX dán lại cùng nhau.

2, Thác triển giải tích

Khi thác triển giải tích dọc theo hai cung khác nhau có thể dẫn đến các kết quả khác nhau mặc dù hai cung có cùng điểm đầu và cuối. Mặt Riemann cũng tham gia vào giải quyết hoàn cảnh này. Bạn nào đọc bài này thì comment giúp cụ thể cái này thế nào nhé! Hôm nay mỏi quá rồi, tạm thế đã!

Thấy bài này hay trên Notices của  AMS

genus

Problem 1.1. (a) We have \infty\not\in\emptyset and \emptyset be open set in \mathbb{R}^n hence \emptyset be open set in X, we have also that X be open in X because \infty \in X and X-X=\emptyset be compact set in \mathbb{R}^n.
Assume that U,V are open sets in X, if \infty\not\in U or \infty\not\in V then \infty\not\in U\cap V and U\cap V is an open set in \mathbb{R}^n therefore U\cap V is open set in X. If \infty\in U and \infty\in V then U=\{\infty\}\cup U',V=\{\infty\}\cup V', here \infty\not\in U'\cup V' and \mathbb{R}^n-U',\mathbb{R}^n-V' are compact sets in \mathbb{R}^n. We have X-(U\cap V)=\mathbb{R}^n-(U'\cap V')=(\mathbb{R}^n-U')\cup (\mathbb{R}^n-V') is compact set in \mathbb{R}^n and \infty\in U\cap V therefore U\cap V be open set in X.
Assume that \{U_i\}_{i\in I} is a family of open sets in X and \infty\in U_i\forall i \in J,\infty\not\in U_i\forall i\in I-J for some J\subseteq I (maybe empty). If \infty\not\in\cup_{i\in I}U_i then \cup_{i\in I}U_i be open in \mathbb{R}^n, and so open in X. If \infty\in\cup_{i\in I}U_i then X-\cup_{i\in I}U_i=[\cap_{i\in J}(\mathbb{R}^n-U_i)]\cap [\cap_{i\in I-J}(\mathbb{R}^n-(U_i-\{\infty\}))] is closed and bound in \mathbb{R}^n, also it is compact set in \mathbb{R}^n, therefore \cup_{i\in I}U_i is open set in X.
Thanks to all conditions above we have X is a topological space.
Assume x\not=y are elements in X. If \infty \not\in \{x,y\} then x,y\in\mathbb{R}^n, note that because \mathbb{R}^n is Hausdorff space we can find open sets U_x,U_y in \mathbb{R}^n such that x\in U_x,y\in U_y and U_x\cap U_y=\emptyset(Note: U_x and U_y are also open sets in X). If \infty\in \{x,y\} , assume that y=\infty, denote U=S(x;1)=\{z\in\mathbb{R}^n|||z-x||<1\},\overline{U}=\{z\in\mathbb{R}^n|||z-x||\leq 1\} and V=\{\infty \}\cup (\mathbb{R}^n-\overline{U}). Then U,V are open sets in X, x\in U,y\in V and U\cap V=\emptyset.
Compactness of X is a immediately follows from (b).
(b)We have \sigma be a surjective because \sigma (0,\cdots,0,1)=\infty and \sigma (x_1,x_2,\cdots,x_{n+1})=(y_1,y_2,\cdots,y_n)\forall (y_i)\in\mathbb{R}^n, where x_i=\dfrac{2y_i}{k+1}(i=\overline{1,n}),x_{n+1}=\dfrac{k-1}{k+1},k=\sum_{i=1}^ny_i^2. But it is easy to see that \sigma is an injective(checking \sigma (x)=\sigma (y) then x=y), therefore \sigma is a bijective. By above formular, we have \sigma is a homeomorphism.(x_n\to x_0\Leftrightarrow \sigma (x_n)\to\sigma (x_0) in topology of \mathbb{R}^n)

Problem 1.2. We have D=\{z\in\mathbb{C}|cz+d\not =0\} is an open set in \mathbb{C} and f'(z)=\dfrac{ad-bc}{(cz+d)^2}\forall z\in D therefore f is holomorphic on D.
If c=0 then ad\not=0 we definition f:\mathbb{P}^1\to\mathbb{P}^1 by setting f(\infty)=\infty.
If c\not =0 then we definition f:\mathbb{P}^1\to\mathbb{P}^1 by setting f(\infty)=\dfrac{a}{c} and f(-\dfrac{d}{c})=\infty.
In both cases we have f is continuous over \mathbb{P}^1 and f is a bijective, it is easy to find formular of f^{-1} and check f and f^{-1} are holomorphic.

Problem 1.3.

Problem 1.4. The problem is wrong, an example \omega_1=1,\omega_2=i and \omega_1'=\sqrt{2}+i,\omega_2'=1+i\sqrt{2} and A=\left(\begin{matrix}\sqrt{2}&1\\1&\sqrt{2}\end{matrix}\right). Then \left(\begin{matrix}\omega_1'\\ \omega_2'\end{matrix}\right)=A\left(\begin{matrix}\omega_1\\ \omega_2\end{matrix}\right) and \omega_1'\not\in \Gamma . In my opinion, A must is belong to \text{SL}(2;\mathbb{Z}).
Problem 1.5. (a)Assume that U is open in \mathbb{C}/\Gamma' and f is original map. We have \pi^{-1}(f^{-1}(U))=\{y\in\mathbb{C}|\pi (y)\in f^{-1}(U)\}=\{y\in\mathbb{C}|\pi'(\alpha y)\in U\} and \pi'^{-1}(U) is open in \mathbb{C} therefore \pi^{-1}(f^{-1}(U)) is open in \mathbb{C}, so $f^{-1}(U)$ is open in \mathbb{C}/\Gamma , or f is continuous.
If \psi_1:U_1\to V_1 and \psi_2:U_2\to V_2 are complex charts of \mathbb{C}/\Gamma and \mathbb{C}/\Gamma', respectively, such that f(U_1)\subset U_2. We have \psi_2\cdot f\cdot \psi_1^{-1}(z)=\psi_2\cdot f ([z]_\Gamma ) =\psi_2 ([\alpha z]_{\Gamma'})=\alpha z\forall z\in V_1 therefore f is holomorphic.
If f is biholomorphic then f must is bijection and \alpha \Gamma =\Gamma'. In fact, if x\in \Gamma'-\alpha \Gamma then f([\dfrac{1}{\alpha}x])=f([0]) but [\dfrac{1}{\alpha}x]\not =[0], contradiction!
If \alpha \Gamma =\Gamma' then f is bijective and \dfrac{1}{\alpha}\Gamma'\subseteq \Gamma therefore by above, f and f^{-1} are holomorphic.
(b)We have \Gamma =\mathbb{Z}\omega_1+\mathbb{Z}\omega_2=\omega_1(\mathbb{Z}+\mathbb{Z}\cdot\dfrac{\pm\omega_2}{\omega_1}) and use (a).
(c)Use (a) again!

Lectures on Riemann Surfaces (Graduate Texts in Mathematics)

by Otto Forster (Author), Bruce Gilligan (Translator) . That is!

Link download http://ifile.it/fyk9zep

 

November 2009
M T W T F S S
« Oct    
 1
2345678
9101112131415
16171819202122
23242526272829
30  

Archives