I have to say I disagree with you.

來源: yma16 2011-01-17 09:51:33 [] [博客] [舊帖] [給我悄悄話] 本文已被閱讀: 次 (933 bytes)
回答: 回複:回複:Yes,you're right.jinjing2011-01-16 21:05:36
Consider sequence A1=a, An+1=a^An for a>0.  If the limit x exists, we can take the limit on both sides and will get x=a^x, which is lna=ln(x)/x.  For a is in (0,1), the limit is the solution of lna=ln(x)/x, which is less than 1.
ln(x)/x is 0 at x=1 and x->infinity.  The max value occurs at e^(1/e). 
When a is in (1, e^(1/e)),
lna=ln(x)/x has two solutions.  We need to use the one that is less than e^(1/e) as the limit.
An example is a=2^.5.  (1)An is increase. (2)An<2
Proof of (2)
A1=a<2
Assume An<2 then An+1=a^An <a^2=2.
Therefore An<2 for all n.
Since (1)An is increase. (2)An<2, An has a limit.  Solve ln(2^.5)=ln(x)/x, we get x=2 or x=4. Accept 2 and reject 4 since it is bigger than e^(1/e).

所有跟帖: 

Thanks,but you are...,Newton was not clear,...100 more years Cau -jinjing- 給 jinjing 發送悄悄話 (287 bytes) () 01/17/2011 postreply 11:12:12

When I can't your thanks, I check my answer again,my god...... -jinjing- 給 jinjing 發送悄悄話 (846 bytes) () 01/18/2011 postreply 12:31:15

回複:When I can't your thanks, I check my answer again,my god..... -yma16- 給 yma16 發送悄悄話 yma16 的博客首頁 (548 bytes) () 01/18/2011 postreply 14:22:39

My friend,you'r great,a>e^(1/e) An is divergent. Why?Because... -jinjing- 給 jinjing 發送悄悄話 (127 bytes) () 01/18/2011 postreply 17:57:16

thank you for the effort. This is the only -yma16- 給 yma16 發送悄悄話 yma16 的博客首頁 (450 bytes) () 01/20/2011 postreply 18:27:29

Thanks,you are right. A old want to say....why the big one is no -jinjing- 給 jinjing 發送悄悄話 (302 bytes) () 01/21/2011 postreply 10:53:30

I also found another example that taking a limit -yma16- 給 yma16 發送悄悄話 yma16 的博客首頁 (373 bytes) () 01/21/2011 postreply 17:42:54

請您先登陸,再發跟帖!

發現Adblock插件

如要繼續瀏覽
請支持本站 請務必在本站關閉/移除任何Adblock

關閉Adblock後 請點擊

請參考如何關閉Adblock/Adblock plus

安裝Adblock plus用戶請點擊瀏覽器圖標
選擇“Disable on www.wenxuecity.com”

安裝Adblock用戶請點擊圖標
選擇“don't run on pages on this domain”