9/20/2013

[集合論] 基礎集合論的數學語言 (3)- Monotone Sequence of Sets

延續前篇  [集合論] 基礎集合論的數學語言 (2)- Limits of Sets

Definition: 單調集合的數列 (Monotone Sequence of Sets)
令 $\{A_n \}$ 為 一組 sequence of sets,我們說 $\{A_n \}$ 為 montone non-decreasing 若 $A_1 \subset A_2 \subset ... $,我們用 $A_n \uparrow$ 表示  $\{A_n \}$ 為 monotone non-decreasing sets。

另一方面,我們說 $\{A_n \}$ 為 montone non-increasing 若 $A_1 \supset A_2 \supset ... $。我們用 $A_n \downarrow$ 表示  $\{A_n \}$ 為 monotone non-increasing sets。

FACT: 對 Monotone Sequence of Sets 其極限存在。


現在我們看個結果

Theorem:
令 $\{A_n \}$ 為 monotone sequence of sets,我們有
  1. 若 $A_n  \uparrow $ ( 亦即 $\{A_n\} $ 為 monotone non-decreasing) 則 \[\lim_{n \rightarrow \infty}A_n = \bigcup_{n=1}^\infty A_n\]
  2. 若 $A_n  \downarrow $ (亦即 $\{A_n\} $ 為 monotone non-increasing) 則 \[\lim_{n \rightarrow \infty}A_n = \bigcap_{n=1}^\infty A_n\]
Proof
先證 (1):令 $\{A_n \}$ 為 monotone sequence of sets  且  $A_n \uparrow $ ,我們要證\[\mathop {\lim }\limits_{n \to \infty } {A_n} = \bigcup\limits_{n = 1}^\infty  {{A_n}} \]注意到若 集合數列的極限存在等價為
\[\mathop {\lim \sup }\limits_{n \to \infty } {A_n} = \mathop {\lim \inf }\limits_{n \to \infty } {A_n} = \mathop {\lim }\limits_{n \to \infty } {A_n}
\]現在由  $A_n  \uparrow $ 我們知道 $A_j \subset A_{j+1}$ 故
\[\bigcap\limits_{k \ge n}^{} {{A_k}}  = {A_n}
\]現在觀察 $\lim \inf$
\[\mathop {\lim \inf }\limits_{n \to \infty } {A_n}: = \bigcup\limits_{n = 1}^\infty  {\left( {\bigcap\limits_{k \ge n}^{} {{A_k}} } \right)}  = \bigcup\limits_{n = 1}^\infty  {{A_n}}
\]接著我們在觀察 $\lim \sup$  可得
\[\mathop {\lim \sup }\limits_{n \to \infty } {A_n}: = \bigcap\limits_{n = 1}^\infty  {\left( {\bigcup\limits_{k \ge n}^{} {{A_k}} } \right)}  \subset \bigcup\limits_{k \ge 1}^{} {{A_k}}  = \mathop {\lim \inf }\limits_{n \to \infty } {A_n}\]又我們知道
\[\mathop {\lim \inf }\limits_{n \to \infty } {A_n} \subset \mathop {\lim \sup }\limits_{n \to \infty } {A_n}
\]故總合以上可推知
\[\left\{ \begin{array}{l}
\mathop {\lim \sup }\limits_{n \to \infty } {A_n} \subset \mathop {\lim \inf }\limits_{n \to \infty } {A_n}\\
\mathop {\lim \inf }\limits_{n \to \infty } {A_n} \subset \mathop {\lim \sup }\limits_{n \to \infty } {A_n}
\end{array} \right. \Rightarrow \mathop {\lim \sup }\limits_{n \to \infty } {A_n} = \mathop {\lim \inf }\limits_{n \to \infty } {A_n}\]且
\[\mathop {\lim \sup }\limits_{n \to \infty } {A_n} \subset \bigcup\limits_{n = 1}^\infty  {{A_n}}  \subset \mathop {\lim \sup }\limits_{n \to \infty } {A_n} \Rightarrow \bigcup\limits_{n = 1}^\infty  {{A_n}}  = \mathop {\lim \sup }\limits_{n \to \infty } {A_n}\]

接著我們證 (2):令 $\{A_n \}$ 為 monotone sequence of sets  且  $A_n \downarrow $ ,我們要證\[\mathop {\lim }\limits_{n \to \infty } {A_n} = \bigcap\limits_{n = 1}^\infty  {{A_n}} \]

同之前證明(1)的手法,若 集合數列的極限存在等價為
\[\mathop {\lim \sup }\limits_{n \to \infty } {A_n} = \mathop {\lim \inf }\limits_{n \to \infty } {A_n} = \mathop {\lim }\limits_{n \to \infty } {A_n}
\] 故現在由定義 $\lim \sup$ 與 $\lim \inf$可知
\[\left\{ \begin{array}{l}
\mathop {\lim \inf }\limits_{n \to \infty } {A_n}: = \bigcup\limits_{n = 1}^\infty  {\left( {\bigcap\limits_{k \ge n}^{} {{A_k}} } \right)} \\
\mathop {\lim \sup }\limits_{n \to \infty } {A_n}: = \bigcap\limits_{n = 1}^\infty  {\left( {\bigcup\limits_{k \ge n}^{} {{A_k}} } \right)}
\end{array} \right.\]由於 $\{A_n \} \downarrow$,亦即 $A_{n+1} \subset A_n$ 故上述  $\lim \sup$ 與 $\lim \inf$ 有如下關係
\[\begin{array}{l}
\left\{ \begin{array}{l}
\mathop {\lim \inf }\limits_{n \to \infty } {A_n}: = \bigcup\limits_{n = 1}^\infty  {\left( {\bigcap\limits_{k \ge n}^{} {{A_k}} } \right)} \\
\mathop {\lim \sup }\limits_{n \to \infty } {A_n}: = \bigcap\limits_{n = 1}^\infty  {\left( {\bigcup\limits_{k \ge n}^{} {{A_k}} } \right) = \bigcap\limits_{n = 1}^\infty  {{A_n}} }
\end{array} \right.\\
\begin{array}{*{20}{c}}
{}&{}
\end{array} \Rightarrow \left\{ \begin{array}{l}
\mathop {\lim \inf }\limits_{n \to \infty } {A_n} \subset \mathop {\lim \sup }\limits_{n \to \infty } {A_n}\\
\mathop {\lim \sup }\limits_{n \to \infty } {A_n} = \bigcap\limits_{n = 1}^\infty  {{A_n}}  \subset \bigcup\limits_{n = 1}^\infty  {\left( {\bigcap\limits_{k \ge n}^{} {{A_k}} } \right) = \mathop {\lim \inf }\limits_{n \to \infty } {A_n}}
\end{array} \right.\\
\begin{array}{*{20}{c}}
{}&{}
\end{array} \Rightarrow \mathop {\lim }\limits_{n \to \infty } {A_n} = \mathop {\lim \inf }\limits_{n \to \infty } {A_n} = \mathop {\lim \sup }\limits_{n \to \infty } {A_n} = \bigcap\limits_{n = 1}^\infty  {{A_n}} \ \ \ \ \ \square
\end{array}\]


上述定理有個重要的結果值得紀錄。
若現在我們取 $B_n$ 為任意 sequence of sets,則
\[\left\{ \begin{array}{l}
\mathop {\inf }\limits_{k \ge n} {B_n}: = \bigcap\limits_{k \ge n}^{} {{B_k}}  \uparrow \\
\mathop {\sup }\limits_{k \ge n} {B_n}: = \bigcup\limits_{k \ge n}^{} {{B_k}}  \downarrow
\end{array} \right.\]利用上述定理可得
\[\left\{ \begin{array}{l}
\mathop {\inf }\limits_{k \ge n} {B_n}: = \bigcap\limits_{k \ge n}^{} {{B_k}}  \uparrow \begin{array}{*{20}{c}}
{}
\end{array} \Rightarrow \begin{array}{*{20}{c}}
{}
\end{array}\mathop {\lim }\limits_{n \to \infty } \left( {\bigcap\limits_{k \ge n}^{} {{B_k}} } \right) = \bigcup\limits_{n = 1}^\infty  {\left( {\bigcap\limits_{k \ge n}^{} {{B_k}} } \right) = \mathop {\lim \inf }\limits_{n \to \infty } {B_n}} \\
\mathop {\sup }\limits_{k \ge n} {B_n}: = \bigcup\limits_{k \ge n}^{} {{B_k}}  \downarrow \begin{array}{*{20}{c}}
{}
\end{array} \Rightarrow \begin{array}{*{20}{c}}
{}
\end{array}\mathop {\lim }\limits_{n \to \infty } \left( {\bigcup\limits_{k \ge n}^{} {{B_k}} } \right) = \bigcap\limits_{n = 1}^\infty  {\left( {\bigcup\limits_{k \ge n}^{} {{B_k}} } \right) = \mathop {\lim \sup }\limits_{n \to \infty } {B_n}}
\end{array} \right.\]

讀者可嘗試以下幾個例子看看是否能正確使用上述定理

Exercise:
\[\begin{array}{l}
\mathop {\lim }\limits_{n \to \infty } \left[ {0,1 - \frac{1}{n}} \right] = [0,1)\\
\mathop {\lim }\limits_{n \to \infty } \left[ {0,1 - \frac{1}{n}} \right) = [0,1)\\
\mathop {\lim }\limits_{n \to \infty } \left[ {0,1 + \frac{1}{n}} \right] = [0,1]\\
\mathop {\lim }\limits_{n \to \infty } \left[ {0,1 + \frac{1}{n}} \right) = [0,1]
\end{array}\]

9/18/2013

[轉載] 如何讓學生真正有熱情與動機「學習」?

如何讓學生真正有熱情與動機「學習」?.
這是一個非常棒的講座,講者 葉丙成老師 與大家分享他在 台大/密大 的個人的教學經驗


講者 葉丙成老師- 台大電機系副教授

一位非常熱血又傑出的老師,分享一個熱血又傑出的演講,當中不論是老師自身準備的Power Point或者教學觀點,理念與想法,很多部分都值得我們深思與學習。老師他分享了多年的教學經驗與引導學生如何從被動中學習變成樂於學習,主動學習
learn by student, assign by student, and grade by student. !

另外 葉丙成老師也有在 Coursera 開授 機率 的 開放式課程 有興趣的朋友可以前往學習



[數學分析] 連續函數族的逐點上包絡函數不一定連續

連續函數有諸多用途,一般在參數最佳化領域中常見的情況是考慮所謂的 上包絡函數(upper envelope function)。 Definition:  定義函數族 \(\{f_t : t \in T\} \) 其中 \(T\) 為 index set 並考慮對任意 \(x ...