Detection Method - GoldenSection.Search()

Oct 5, 2011 at 1:27 PM

Hello,

While developing, I noticed that it was kind of hard for some people to recreate a posture previously recorded. I decided to try to tune the detection precision down, so I went through the code, and arrived finally at the method GoldenSection.Search(...). There, I noticed the use of certain variables, like ReductionFactor and Diagonal, and, well, I was confused. What are their importance to the code? I grasped some of the logic, but a lot of it stills elude me, like the return value:

return 1.0f - 2.0f * min / Diagonal;
 
I don't know if I'm analyzing the right method, and if there is any documentation on this, but, what sould I do to lower the detection precision?

Thanks a lot!

Coordinator
Oct 5, 2011 at 2:03 PM

Hi,

you can find a description of the GoldenSection Search algorithm:

http://www.math.uic.edu/~jan/mcs471/Lec9/gss.pdf

Hope this helps :)

Jun 24, 2014 at 10:45 PM
Hi, I have studied the pdf for a method to locate the minimum of a function, however I still don't understand the use of the rotate method within search.
Even before the distance between sequences has been calculated the array rotatedList is calculated using current.Rotate(x1) like this:

List<Vector2> rotatedList = current.Rotate(x1);

Why do you keep rotating the current sequence and calculating the distance until a and b are very close?, the function which we want to minimize has as its argument an angle and as output the distance?

Any help will be much appreciated