作者ppia (papayaPaul)
看板Math
標題Re: [分析] 實變
時間Sat May 15 23:24:39 2010
※ 引述《smartlwj (還有3間)》之銘言:
: 1. Let μ be a measure (on some measure space X) and let f_n, n=1,2,...
: be a sequence of real-valued measurable functions on X. Suppose
: that ofr every ε>0, the sum
: ∞
: Σ μ{x| |f_n(x)|>ε} is finete.
: n=1
: Prove that f_n converges to zero almost everywhere.
: 2. Let I=(0,1). Suppose {f_n} is a norm-bounded sequence of function in L^2(I)
: that converges in measure to a function f.
: (a) Show that f in L^2(I) and ║f║≦ liminf║f_n║
: 2 2
: (b) Show that ║f_n║ converges to ║f║ if and only if ║f_n - f║ → 0.
: 2 2 2
: 請問這兩題該怎麼做呢??
: 第一題我由題目可以知道 f_n 是依測度收斂到0
: 所以存在子列{f_nj}會幾乎處處收斂到0. 然後我想要証明出f_n →0 a.e.
: 由|f_n-0|≦|f_n-f_nj| + |f_nj| 然後我就不知道該怎麼做了
: 第二題的(a)沒想法
: (b)的(<=)已証出
: (=>)我想用LDCT 但是 我缺少了 f_n → f a.e.的條件
: 而題目只有 f_n → f in measure.
: 我該怎麼做呢??
: 不知道我目前的想法是否有問題?請指教 謝謝<(_ _)>
(a) Let f_nk be a subsequence of f_n such that
lim ║f_nk║ = lim inf║f_n║.
k→∞ n→∞
For any ε>0, put J_n = {x│|f(x)|>ε}. Since lim μ(J_n) = 0, there exists a
subesequence J_nki of J_nk such that
∞
Σ μ(J_nki)≦1.
i=1
Put
∞
E_i = I - ∪ J_nks.
s=i
{E_i} is then an increasing sequence of measurable sets with lim μ(E_i) = 1
and E_i ㄈ I-J_nki. Hence
2 2 2 2 2 2
∫ |f| ≦ ∫ |f| ≦ ∫|f_nki-f| + ∫|f_nki| ≦ ε + ║f_nki║
E_i I-J_nki I-J_nki I-J_nki
2
By the monotone convergence theorem, the first term approachs ║f║ as i goes
to infinity.
(b) (=>) For any ε>0, let J_ns be a subsequence of J_n such that
∞
Σ μ(J_ns)≦1.
s=1
Put ∞
C_s = ∪ J_nt and D_s = I- C_s.
t=s+1
2 2 2 2 + 2
∫|f_ns| = ║f_ns║ - ∫ |f_ns| ≦ ║f_ns║ - ∫ [(|f|-ε) ]
C_s D_s D_s
Hence
2 2 + 2
lim sup ∫|f_ns| ≦ ║f║ - ║(|f|-ε)║ .
s→∞ C_s
2 2 2 2 2 2 2
║f_ns-f║ ≦ ∫|f_ns-f| + ∫|f_ns| + ∫|f| ≦ ε + ∫|f_ns| + ∫|f|
D_s C_s C_s C_s C_s
Therefore 2 2 2 + 2
lim sup ║f_ns-f║ ≦ ε + ║f║ - ║(|f|-ε)║
s→∞
Since ε was arbitrary, lim ║f_ns-f║= 0. Now that lim ║f_n║=║f║,
lim ║f_n-f║= 0. s→∞ n→∞
n→∞
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 125.231.213.40
※ 編輯: ppia 來自: 125.231.221.157 (05/16 12:46)
推 math1209 :推…事實上,只要 conv in measure 則可取代逐點收斂 05/16 16:58
→ math1209 :這樣一來, 本題就顯得簡單許多了. 05/16 16:59
推 smartlwj :推一下~感謝p大 05/16 17:08
→ ppia :不懂m大的推文 可以解釋一下嗎? 05/16 18:45
推 math1209 :我的意思是說 舉凡 LDCT, MCT, Fatou's lemma 或者 05/16 19:25
→ math1209 :更廣義的 LDCT. 這些條件中的 幾乎處處逐點收斂 05/16 19:26
→ math1209 :被換成 依測度收斂, 其結論也是成立的. 05/16 19:26
→ ppia :不知道是不是我會錯意 但是 05/16 19:28
→ ppia :fn→f in measure =\=> fn→f a.e. 05/16 19:29
→ ppia :抱歉 我沒看清楚推文...沒事了... 05/16 19:32