台灣大學數學系

九十二學年度第二學期博士班資格考試題

實分析

May 9, 2004

[回上頁]


There are Problems A to D with a total of 110 points.Please write down your proof or computatioinal steps clearly on the answer sheets.

A.
Let $ (X, \mathcal{M}, \mu)$ be a measure space where $ \mathcal{M}$ is a $ \sigma$-algebra of subsets of $ X$, and $ \mu:
\mathcal{M} \rightarrow [0, \infty]$ is a measure. Define

$\displaystyle \mu^*(E) = \inf \{ \mu(A) \vert A \in \mathcal{M}$   , and $\displaystyle E \subset A \}$    , for any subset $\displaystyle E \subset X$

(a)
(10 points) Prove that $ \mu^*$ is an outer measure define in $ X$ such that every $ E \in \mathcal{M}$ is $ \mu^*$-measurable, and $ \mu^*(E) = \mu(E)$. Are sets in $ \mathcal{M}$ the only $ \mu^*$-measurable subset of $ X$?
(b)
(15 points) Let $ E_n \subset X(n = 1, 2, 3, \cdots)$ be a sequence of subsets. Probe that $ \mu^* (\lim \inf_{n \rightarrow \infty} E_n) \leq \lim \inf_{n \rightarrow \infty} \mu^* (E_n)$. Give example to show that the equality is in general false. Is $ \mu^*(\lim \sup_{n \rightarrow \infty} E_n)
\geq \lim \sup_{n \rightarrow \infty} \mu^*(E_n)$ also true?
B.
(15 points)Let $ f \in L^1 (\mathbb{R}^n)$ and $ K\subset \mathbb{R}^n$ be a compact subset of $ \mathbb{R}^n$. Define the functions

$\displaystyle \phi (x)= \int_\mathbb{R}^n \vert f(x+y)+f(y)\vert dy, \bigskip\psi(x)=\int_{k+x} f(y)dy\bigskip $    for $\displaystyle \bigskip x \in \mathbb{R}^n$

where $ K+x=\{y+x \vert y \in K \}.$ Prove that $ \phi$ and $ \psi$ are both continuous functions in $ \mathbb{R}^n$.Are both uniformly continuous in $ \mathbb{R}^n$?
C.
Let $ (X, \mathcal{M}, \mu)$ be a measure space ( as in Problem A). Let $ f_n( n = 1, 2, 3, \cdots )$ and $ f$ be in $ L^p(X, \mathcal{M}, \mu)$ with $ 1 \leq p < \infty$.
(a)
(15 points) If $ \lim_{n \rightarrow \infty} \vert\vert f_n\vert\vert _{L^p} = \vert\vert f \vert\vert _{L^p}$, prove that $ f_n \rightarrow f$ in measure iff $ \lim_{n \rightarrow \infty} \vert\vert f_n - f \vert\vert _{L^p} = 0$
(b)
(15 points) If $ \vert\vert f_n\vert\vert _{L^p}$ is bounded in $ L^p$ and $ f_n \rightarrow f$ in measure where $ 1 < p < \infty$, prove that

$\displaystyle \lim\limits_{n \rightarrow \infty} \int_X f_n(x) g(x) d \mu(x) = \int_X f(x) g(x) d\mu(x)$    for any $\displaystyle g \in L^q(X, \mathcal{M}, \mu)$

where $ q$ is given by $ \frac{1}{p} + \frac{1}{q} = 1$. Show that by an example that the conclusion is false in case $ p = 1$.
D.
Determine which of the following statements is true or false. Prove your answer. Each has 10 points.
(a)
Let $ f:[a,b] \rightarrow \mathbb{R}$ be a function of bounded variation. Then $ f$ is absolutely continuous in $ [a,b]$ iff $ f$ is absolutely continuous in $ [a+\varepsilon, b]$ for all sufficiently small $ \varepsilon > 0$.
(b)
Let $ E \subset \mathbb{R}^2$ be a Borel subset, then the set $ F = \{ \sin(x^2 + y^2) \vert (x,y) \in E \} $ is also Borel in $ \mathbb{R}$. However, if $ E$ is Lebesgue measurable, the $ F$ may not be Lebesgue measurable.
(c)
There exists a collection $ \mathcal{F}$ of closed rectangles in $ \mathbb{R}^n$ such that $ \cup_{E \in \mathcal{F}} E$ is not Lebesgue measurable.
(d)
Let $ f : \mathbb{R}^2 \rightarrow \mathbb{R}$ be Borel measurable. Then, for each $ y \in \mathbb{R}$ fixed, the function $ f(x,y)$ is also Borel measurable in $ x \in \mathbb{R}$.

[回上頁]