hi..
i'm a student here...
having a problem in programming question..
hope u all cal help me...plz..

Question:

1)An integer number is said to be a perfect number if its factors, including 1 (but not the number itself), sum to the number. For example, 6 is perfect number, because 6 = 1+2+3. Write a method Perfect that determines whether parameter number is a perfect number. Use this method to display all perfect numbers between 1 and 100.


2)Write a method Multiple that determines, for a pair of integers, whether the second integer is a multiple of the first. The method should take two integer arguments and return true if the second is a multiple of the first and false otherwise.
Thanks for helping!

Rasit, the simple version of simplex (i dont about its variants) relies on linearity of boundary of feasible region, while in SDP, these boundaries are non-linear which makes the problem difficult. Still they are only piecewise algebraic surfaces and not general non-linear surfaces so the methods for solving SDP try to exploit this fact in some way.

Edited by author 10-14-2004 02:38 PM
I wonder why can we not extend the Simplex method to solve for semidefinite problems in a straightforward way.

Robin, I couldn't find any paper with title "subgraph matching for computer vision" although there were several related papers. Which paper will you cover example from ? Can you post a link for it ?

If in some application, after formulating the problem as SDP we end up with many constraints and we dont want to force our solution to satisfy all of them, e.g. because some of the constraints may be outliers. How are such situations handled in practice ?

A specific application example would be nice to bring everything together

Interesting. Doing a quick google search of semidefinite programming and computer vision, I found that it's been used in various applications, including image segmentation, preceptual grouping, camera calibration, and some unsupervised learning for image apps.

The paper (/book? ;) ) describes several applications of SDP to control theory and combinatorial optimization by listing several examples. My question is, what types of problems in Computer Vision is SDP used to solve?

Gary, thanks for asking! I'll focus mainly on setting up, rather than solving an SDP problem. Time permitting, however, I'd like to go thru a high-level walkthru of how SDPs are solved and leave you with a roadmap thru the long and winding discussion in sections 3-5.

So, I recommend reading sections 1 & 8 carefully. For section 2, I'd suggest looking thru the examples, but not getting bogged down in things like Young's modulus of elasticity. The main example I'll cover is from a different paper - subgraph matching for computer vision. Also, I'll go over the example of handling a nonlinear constraint (pg 3) using Schur complements. Then, if you can hang tough to read sections 3-5 thru once that will help you get more out of the walkthru I hope to give. I won't cover anything from sections 6 or 7.

Hope this helps,
Robin

This review seems quite interesting, but is bit of a monster in size. Can you suggest particular areas/sections of the paper I should concentrate on first.

Thanks.