Dear all!
I have a problem to choose a good numerical method
for the following problem.Does somebody have an idea?
I have to solve the equation U(theta)=0 for theta=(beta, sigma, gamma)-
parameter vector. U is a multidimensional function of theta.
But a problem is that the surface is very outspread (flat), so it's
difficult sometimes to find the solution. As a result, Newton-Raphson
sometimes doesn't converge and I should to stop the process and to
simulate new data set. I am using now Newton Raphson with step factor.
What may be an appropriate idea to improve it? May be we have such
algorithms on S? C?
Thank you in advice, Viki.
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|