# [ODE] Good reference for solving LCP's?

Michael Goldish michael.goldish at gmail.com
Wed Apr 25 04:22:18 MST 2007

```Here's a classical paper by Baraff which discusses the pivotal algorithm
implemented in LCP.cpp:
http://www.cs.cmu.edu/~baraff/papers/sig94.pdf

and here's a paper which can help you understand the iterative algorithm in
stepfast.cpp:
http://www.continuousphysics.com/ftp/pub/test/index.php?dir=physics/papers/&file=IterativeDynamics.pdf&

Why are you saying the constraint style does not form a LCP? As far as I can
tell, the algorithm in LCP.cpp is an implementation of the classical LCP
solver introduced by Baraff. Joints move between the groups C (clampled) and
NC (unclamped - broken) during algorithm runtime.

On 4/24/07, Nguyen Binh <ngbinh at gmail.com> wrote:
>
> Actually, ODE's constraint style does not form a LCP. It's just an
> constrained-ODE (Ordinary Differential Equation) .
> There are some other constrained model that really form a LCP
> (Stewart-Trinkle , Anitescu-Potra time stepper).
>
> The key difference is with LCP model, the contact can break in-between a
> time step not like ODE-style where contact can only break at the end of the
> time step.
>
> On 4/24/07, Jon Watte (ODE) <hplus-ode at mindcontrol.org> wrote:
> >
> > Well, ODE does more than just solve ODEs. It solves the system of joint
> > constraints, which is where the LCP solver comes in. One google term
> > might be "big matrix solver" or "constraint relaxation" if you don't
> > want to just plug in LCP into MathWorld.
> >
> > Cheers,
> >
> >           / h+
> >
> >
> > Megan Fox wrote:
> > > I realize it's a generic response, but MathWorld is an excellent
> > > jump-off point for anything and everything math related.  Wikipedia is
> >
> > > a close second, once you have a point of reference to start from.
> > >
> > >
> > > ... or more generally, numerical solutions to Ordinary Differential
> > > Equations (or ODE's, hence the name).  I'd suggest a textbook about
> > > Numerical Analysis period, but... mine is terrible, and everyone else
> > > I know has a terrible one too.  If you find a decent one, let me know,
> > > mine is a glorified paper-weight occasionally useful in holding down
> > > my solid gold class notes.
> > >
> > > On 4/24/07, Andrew Riehm <andrew.riehm at gmail.com> wrote:
> > >
> > >> I'm trying to understand how the physics engine works, from a
> > numerical
> > >> computing perspective, and I am having a hard time finding a good
> > >> comprehensive overview of what a linear complimentarity problem is
> > and
> > >> what methods are used to solve them.  Can anyone point me to a good
> > >> reference (be it an online article, a book, or a journal)?
> > >>
> > >> --
> > >> Andrew Riehm
> > >> _______________________________________________
> > >> ODE mailing list
> > >> ODE at ode.org
> > >> http://ode.org/mailman/listinfo/ode
> > >>
> > >>
> > >
> > >
> > >
> > _______________________________________________
> > ODE mailing list
> > ODE at ode.org
> > http://ode.org/mailman/listinfo/ode
> >
>
>
>
> --
> --------------------------------------------------
> Binh Nguyen
> Computer Science Department
> Rensselaer Polytechnic Institute
> Troy, NY, 12180
> --------------------------------------------------
>
> _______________________________________________
> ODE mailing list
> ODE at ode.org
> http://ode.org/mailman/listinfo/ode
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ode.org/pipermail/ode/attachments/20070425/fbcc057d/attachment.htm
```