Smallest simplex
From Wikimization
| Line 10: | Line 10: | ||
The question is: can this problem be cast into a convex optimization? | The question is: can this problem be cast into a convex optimization? | ||
| - | Doesn't a simplex in n-space always have (n+1) vertices (http://en.wikipedia.org/wiki/Simplex)? Or would you want to allow for sub-dimensional simplices? But then, measuring the volume is quite pointless for all but the full-dimensional ones. I'm probably misunderstanding something, perhaps you can clarify this? | + | ''Doesn't a simplex in n-space always have (n+1) vertices (http://en.wikipedia.org/wiki/Simplex)? Or would you want to allow for sub-dimensional simplices? But then, measuring the volume is quite pointless for all but the full-dimensional ones. I'm probably misunderstanding something, perhaps you can clarify this?'' |
To contact me, my E-mail address is: yuyue05@mails.tsinghua.edu.cn. | To contact me, my E-mail address is: yuyue05@mails.tsinghua.edu.cn. | ||
Revision as of 06:34, 30 June 2008
I am a PhD. candidate student in Tsinghua University, China. I think this is an open problem in my field. That is:
How to find the smallest simplex which can enclose a bunch of given points in a high dimensional space (under the following two assumptions:)?
- (1) The number of the vertexes of the simplex is known, say n;
- (2) The number of the vertexes of the simplex is unknown.
To measure how small the simplex is, we can use the volume of the simplex.
The question is: can this problem be cast into a convex optimization?
Doesn't a simplex in n-space always have (n+1) vertices (http://en.wikipedia.org/wiki/Simplex)? Or would you want to allow for sub-dimensional simplices? But then, measuring the volume is quite pointless for all but the full-dimensional ones. I'm probably misunderstanding something, perhaps you can clarify this?
To contact me, my E-mail address is: yuyue05@mails.tsinghua.edu.cn.