Let {fn} be a sequence of meassurable functions which converges a.e on [0,1] to a bounded function f.show that there exists pairwise disjoint measurable sets {An}n=0∞with union [o,1]such that A0 has measure zero and fn uniformly on each Ak,k≧1這題要用Egorov's Theorem 下去作:Geven any ε>0,For any k€N,choose Ek包含於[0,1],such that m([0,1]\\Ek)<1/k(這是我預設的),and choose Nk,snch that |fn(x)-f(x)|<ε,for all x€Ek,for all n€NkNow let Ak=Ek∩?到這裡就卡住了,要如何造出一兩兩不相交的區間,使得它們的聯集是[0,1] 或者是(0,1),有哪位高手可以完成證明?
2006-10-27 15:03:47 · 2 個解答 · 發問者 ? 7 in 科學 ➔ 數學
前面的論證應該是:Geven any ε>0,for any k€N,choose closed set Ek,such that m([0,1]\Ek)<ε(這部分到時候要作適度調整),and choose Nk,such that |fn(x)-f(x)|<1/k,for all x€Ek and for all n≧Nk
2006-10-27 15:41:33 · update #1
Egorov's Theorem => there exist A_k in Σ such that μ([0,1]-A_k) < 1/k and f_n -> f uniformly on A_k for each k = 1,2,...Take A = ∪_[1,oo] A_k, then μ([0,1] - A) ≦ μ([0,1] - A_k) < 1/k for all k, this implies that μ([0,1] - A) = 0.May write ∪_[1,oo] A_k = ∪_[1,oo] B_k, where B_k's are disjoint and B_k is contained in A_k, then f_n -> f uniformly on each B_k obviously, so we complete the proof.
2006-10-28 19:56:58 補充:
A 不一定是[0,1],但只比[0,1]小一點點, 因為 μ([0,1] - A) = 0. 扣掉這個 A 不看, 剩下來的可以寫成B_k 的聯集, 而這些 B_k 是互相disjoint的, 而且 f_n 會在每個 B_k 上均勻收斂
2006-10-28 19:58:29 補充:
打太快, 應該是扣掉這個 [0,1]-A 不看, 剩下來的 (就是A) 可以寫成 ...
2006-10-28 20:13:30 補充:
[0,1]不一定會等於 A 阿
你題目不是要證存在互不相交的 A_0,A_1,...使得這些集合聯集起來是[0,1]且有μ(A_0)=0,且 f_n 在每個 A_k, k=1,2,... 上都均勻收斂 ?
2006-10-28 20:14:36 補充:
A_0 就是我的 [0,1]-A
A_k 就是我的 B_k, k=1,2,...
2006-10-28 20:30:35 補充:
NO,
B_1 = A_1
B_k =
∪(1,k)A_k - ∪(1,k-1) A_k
for k = 2,3,...
2006-10-28 22:29:22 補充:
folland 的實變
p61. exercise.41 有更廣的說法
If μ is σ-finite and f_n -> f a.e., there exist measurable E_1,E_2,... are all contained in X such that μ((∪E_k)^c) = 0 and f_n -> f uniformly on each E_k.
2006-10-28 23:02:16 補充:
5本吧, 有三本沒什在看 ...
的確是越多本越好
2006-10-28 23:14:36 補充:
我在看是還OK
2006-10-30 22:32:01 補充:
原來你在用 Zygmund 的書喔 @@
剛剛在從實變中找機率的東西無意中翻到 ...
2006-10-31 09:05:29 補充:
那本我有印阿 只不過沒什在看
只拿來查東西而已 後面習題翻到的
因為它 LEBESGUE 的積分定義和別本不太一樣
2006-10-27 19:07:55 · answer #1 · answered by L 7 · 0⤊ 0⤋
天峨,請問你是用[0,1]\A的測度為零,來證明A=[0,1]嗎?
2006-10-28 20:02:50 補充:
那要如何知道[0,1]=A=∪(k=1,∞)Ak
2006-10-28 20:20:01 補充:
你∪(1,∞)Ak=∪(1,∞)Bk是利用σ-algebra的性質?
2006-10-28 20:41:12 補充:
之前有作過類似的題目嗎?
2006-10-28 22:36:07 補充:
哇!你有幾本實變的書,我目前手上只有兩本,是不是越多本越好?
2006-10-28 23:12:42 補充:
這樣的話,不是會越看越亂,因為每本書的定義和符號都不相同
2006-10-30 23:15:52 補充:
你在那裡翻到?
2006-10-28 15:07:31 · answer #2 · answered by ? 7 · 0⤊ 0⤋