Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
Hi tarkyss
There is an itrative method for calculation of maximum aigenvalue of a matrix by name of Method of Power that I think it is easy to implement.
I saw it in the book "Numerical Methods in Electromagnetics" by Sadiko.
Now time is late, if you cannot find any information about this method notify me.
the itrative counter are constant?
if not, i think it may be not for suit for hardware implemented, because the time cannot be controlled,
i searched many algorithms, for example QL algorithm JACOBA algorithm exponent multiplier, but all algorithms have the same question, maybe there are some hadware method to solve this question, but i dont know
hi tarkyss
1- In this method you can only calculate the biggest eigenvalue.
2- Like other iterative algorithms, more iteration lead in better and more exact results.
regards
use newton method to get every eigenvalues, and sorting.
The iteration count ..... it depends on the precision you want and initial value. I think to try it is the best policy
but if we want to restrict the iteration count,
for example, the iteration cannot more than 10 times. newton method cannot tell me the precision for any matrix, so i dont think it is a good method for vlsi implemention
Hi tarkyss
I've attached for you some part of "Numerical Methods in Electromagnetics" by Sadiko.
it explaine both Power Method & Jacobi Method.
as dinolin said number of iteration depends on desired precision. but maybe some methods converge better to the answer.
I think the Power Method has no matter with complex problem.
Best. mah
thanks, but for vlsi implemention, if the maximum iteration number is 10, can you tell me the precision for any matrix? of course the iteration depends on desired presion, but it depends on data too.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.