z*******x 发帖数: 30 | 1 【 以下文字转载自 EE 讨论区 】
发信人: zhuzhufox (I+love+you+), 信区: EE
标 题: Extreme value
发信站: BBS 未名空间站 (Wed Aug 8 13:50:53 2007)
I have an extreme value problem which confused me a lot....
suppose delta=p-q, where p & q are positive and a function of x, and delta
has an maximum value at xc, that is, d(delta)/dx=0 at xc.
Now can I prove that (c-p)/q also has an extreme value at xc??? If yes, how
to prove it???
Thanks a lot advanced!! | z*******x 发帖数: 30 | 2 BTW, c is a constant number . | n*******l 发帖数: 2911 | 3 No.
For axample, let p = -x^2 +x+1 and q = x+1. Then
p-q = -x^2 and it attains max at x=0.
But (c-p)/q = (c+x^2-x-1)/(x+1) goes to infinity as x approaches infinity.
how
【在 z*******x 的大作中提到】 : 【 以下文字转载自 EE 讨论区 】 : 发信人: zhuzhufox (I+love+you+), 信区: EE : 标 题: Extreme value : 发信站: BBS 未名空间站 (Wed Aug 8 13:50:53 2007) : I have an extreme value problem which confused me a lot.... : suppose delta=p-q, where p & q are positive and a function of x, and delta : has an maximum value at xc, that is, d(delta)/dx=0 at xc. : Now can I prove that (c-p)/q also has an extreme value at xc??? If yes, how : to prove it??? : Thanks a lot advanced!!
| A*******r 发帖数: 768 | 4 general fractional programming
NP-hard |
|