全部版块 我的主页
论坛 经济学论坛 三区 微观经济学
11362 13
2014-10-19
用海塞矩阵证明函数凸性时,需要保证海塞矩阵的顺序主子式全部半正定吗?
另外什么文献可以找到海塞矩阵半正定推出函数为凸函数的证明,谢谢
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

全部回复
2014-10-19 22:37:04
If you want to show that a function is convex (not strictly convex), only LEADING principle minors >=0 is not enough. You need to show all principle minors >=0.

Google Hessian and principle minors.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2014-10-20 00:26:20
pingguagain 发表于 2014-10-19 22:37
If you want to show that a function is convex (not strictly convex), only LEADING principle minors > ...
thank you very much and I have another question .let a set D1 is a convex set and D2 is the other one,then is the set D1-D2  also a convex set ? what's more ,how to proof.   
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2014-10-20 00:54:21
【炊烟】 发表于 2014-10-20 00:26
thank you very much and I have another question .let a set D1 is a convex set and D2 is the other  ...
Maybe I have understood the problem.But i am  wondering whether we can use Hessian  matrix to show a function is convex when  independent  variables in a  small interval . For it is too difficult to proof it by using the definition of convex set.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2014-10-20 04:47:34
【炊烟】 发表于 2014-10-20 00:26
thank you very much and I have another question .let a set D1 is a convex set and D2 is the other  ...
The answer is no. Think about two circles and one contains the other one. The difference is not convex any more.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2014-10-20 04:48:58
【炊烟】 发表于 2014-10-20 00:54
Maybe I have understood the problem.But i am  wondering whether we can use Hessian  matrix to show ...
I didn't get your question. What do you mean "independent variables in a small interval". It is not a complete sentence.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

点击查看更多内容…
相关推荐
栏目导航
热门文章
推荐文章

说点什么

分享

扫码加好友,拉您进群
各岗位、行业、专业交流群