Well, I've implemented some explicit integration methods( Euler, MidPoint, Heun and RK4 ). I'm looking for a implicit method and I've read that the symplectic one was good in terms of accuracy and energy conservation. Unfortunately, I haven't read a paper sufficiently general to provide enough insight to take on a more serious problem than the resolution of d^2x/dt^2 = a(t)( x(t) being the vector position at time t, and a(t) the acceleration ) Here's there a good reference book on the symplectic integration method for the motion of a rigid body? Thank you for your help. PS: the system of equation I want to integrate is the one Baraff provides in his paper, "Physically Based Modeling" [Edited by - johnstanp on June 28, 2008 4:08:18 AM]