发新话题
打印

今天上课prof 问的一条数学题

I also used differential, but still have problem....
let X=Sum[i^d, {i, 1, n}]
for n is fixed:
dX = Sum[d i^(d-1) di, {i, 1, n}]
I have to assume not all di s are the same, so only the highest order di survive....
dX = d n^(d-1) di

for i fixed:
dX = d(n^d)= d n^(d-1) dn

so dn = di

but this prove seems not perfect...
引用:
原帖由 pi-po-si 于 2007-4-4 11:31 AM 发表


嘿嘿,d/di 的时候 对所有的 i 求导啊
这个需要理解才能接受,所以偶就家了一个note,不容易说明白。

是个中间过程,对之求导,之后,delt 其实=0
[ 本帖最后由 andrew 于 2007-4-4 11:40 AM 编辑 ]
The best is yet to come

TOP

引用:
原帖由 andrew 于 2007-4-4 11:36 AM 发表
I also used differential, but still have problem....
let X=Sum
for n is fixed:
dX = Sum
I have to assume not all di s are the same, so only the highest order di survive....
dX = d n^(d-1) di
...
sigma 和积分 是同性质操作,类是交换积分顺序,你看看这个是不是能帮助你理解过去

TOP

what is 类 in english?
引用:
原帖由 pi-po-si 于 2007-4-4 11:48 AM 发表


sigma 和积分 是同性质操作,类是交换积分顺序,你看看这个是不是能帮助你理解过去
The best is yet to come

TOP

引用:
原帖由 andrew 于 2007-4-4 11:50 AM 发表
what is 类 in english?

hope it helps

附件

2.jpg (4.79 KB)

2007-4-4 12:56 PM

2.jpg

TOP

引用:
原帖由 pi-po-si 于 2007-4-4 11:56 AM 发表


hope it helps
it doesnt say anything.....
because you used di = dn in the 2nd "="
which is the thing we want to prove.....
The best is yet to come

TOP

whatever

TOP

o, i think the answer may be easier than i think....
because for any i , we can write
i + A =n
where A is a constant
so di = dn

btw, thanks pi-po, your answer is really brilliant!
The best is yet to come

TOP

引用:
原帖由 andrew 于 2007-4-4 12:06 PM 发表
o, i think the answer may be easier than i think....
because for any i , we can write
i + A =n
where A is a constant
so di = dn

btw, thanks pi-po, your answer is really brilliant!
在偶放弃的时候,终于被接受了,看来再坚持一下的格言非常正确,

你的考虑很正确的,在这个证明里,有很多需要说明,比如那个constant,其实有一个back ground cancel, 为什么取,C 不是, c1,c2,c3,c4....cn

因为最后,nc1 +(n-1)c2+(n-2)c3+....
当n=0的时候, c2,c3,c4...=0
所以就取 C1=c2=c3=...=c

这个证明如果写严格的却很麻烦的

TOP

引用:
原帖由 pi-po-si 于 2007-4-4 12:12 PM 发表



在偶放弃的时候,终于被接受了,看来再坚持一下的格言非常正确,

你的考虑很正确的,在这个证明里,有很多需要说明,比如那个constant,其实有一个back ground cancel, 为什么取,C 不是, c1 ...
The best is yet to come

TOP

引用:
原帖由 andrew 于 2007-4-4 12:06 PM 发表
o, i think the answer may be easier than i think....
because for any i , we can write
i + A =n
where A is a constant
so di = dn

现在明白了,难怪刚才偶答非所闻

这个 di/dn 的积分变量变换jaccobbi 行列式 为什么等于1是你的问题

如果 不保证 A=constant,就失去了n和i之间的固定差的性质, 所以 di=dn, dA=0

TOP

发新话题