My Math Forum (http://mymathforum.com/math-forums.php)
-   Linear Algebra (http://mymathforum.com/linear-algebra/)
-   -   Optimization of a Matrix with one constraint. (http://mymathforum.com/linear-algebra/2033-optimization-matrix-one-constraint.html)

 shack December 16th, 2007 07:03 AM

Optimization of a Matrix with one constraint.

Hello!

I have a minor problem, that wont get solved...

Its been posted in an old competition at our university, but I didnt manage to solve it in time. Maybe someone in here could help me?
http://img504.imageshack.us/img504/4...ringts0.th.jpg

Where D is a symmetric n x n matrix, and x is an unknown n x 1 vector.

I can only figure out how to test for a maximum later on, but is there something im missing with the charataristic roots?

Regards

 mathman December 16th, 2007 05:56 PM

I must confess I'm a little rusty here. However by taking first derivatives with respect to each of the xi, you will end with a matrix equation with (x1,....,xn) any eigenvector, as a solution. Normalize the eigenvectors to unit length and see which one gives you the maximum.

 shack December 16th, 2007 11:37 PM

You dont sound rusty at all in this, compared to me that is.

We tried just to solve it for all x's and came to the following:

http://img225.imageshack.us/img225/1...trixyi9.th.jpg

But we can't eliminate lamda. Do we need to diagonalise D before doing so?

 mathman December 17th, 2007 02:11 PM

If you look at the line which says "D is symmetric" and you make the equations into one big vector, you will have Dx=Lx (L=lambda), which means the solution is an eigenvector and L is an eigenvalue. In general there will be n solutions.

 All times are GMT -8. The time now is 12:18 AM.