fred@inuxc.UUCP (06/18/83)
Well after spending an afternoon working out the math it seems that a 4th order Runga Kutta numeric integration is exactly equivalent to.... x=xo+vo*t+.5*a*t**2 for n body problem, point gravity, no drag. That explains why Lew had reasonable results using it in college. The error using this form will be of the order of... fifth derivative of position*(time step)**5 a good approximation. It also means that this method should be good enough to start a predictor- corrector integration instead of having to go through the pain of starting it with R-K. I'm off fishing for a week. Fred (ABI till August)