28 May
2006
28 May
'06
11:47 a.m.
William Swanson wrote:
Using an infinitely small dt would give you exact integration, but computers can't do that. :) Using something called Simpson's rule, however, can improve the results dramatically with almost no extra work:
IIRC "Add Nodes" uses an implementation of Simpson's rule. I also tried a few other methods, but Simpson's rule was for sure the fastest. I think njh would tell you that eliptic integrals will give a much better/faster approximation, but that math was way over my head.
Aaron Spike