[ODE] Euler vs. Runge-Kutta and adaptive step sizes

Gary R. Van Sickle g.r.vansickle at worldnet.att.net
Tue Apr 30 21:49:01 2002


> I agree with you that ODE's LCP solver is also integrated with its
> integrator to some degree, but this dependency can be easily removed (though
> I would not encourage doing that).
>

I would! ;-)  I'm currently modelling a system which goes from unusably unstable
(in the its-only-oscillating/thank-God-for-ODE sense) to rock-solid if I double
the framerate.  Unless my poker knowledge fails me, one trapezoid beats a pair
of Eulers by a pretty good margin.

Of course, the implementation details may be more non-trivial than one would
suspect ;-).

--
Gary R. Van Sickle
Brewer.  Patriot.