Runge-Kutta Method

Hi I've been reading about the Runge-kutta method to numerically solve diferential ecuations, I understand the concept, although my maths are pretty rusty, but I dont know how I can implement this in Vb.Net.

I've checked codes in VB and other languages like Java, or C#, but each time they use runge-kutta they just call a runge-kutaa function, and I cant access it.

So basically my question is how is one meant to code this method. I feel like it shouldnt be so hard, but right now Im not on the top of my game, I've been all day in front of the computer.

It is for coding spring behaviour, but in general it should be very usefull to solve any type of diferential ecuations.

I know kangaroo does springs, but Im trying to get to code it. I am really intereted in coding physical behaviour, although I know VB. Net is probably not the best for this, but it is all I know as of now.

So any help would be much apreciated.

Thanks in advance for any tips you might be able to give.

Replies to This Discussion

Permalink Reply by Daniel Piker on December 3, 2011 at 3:54pm

http://gafferongames.com/game-physics/integration-basics/

Permalink Reply by Nicolas on December 5, 2011 at 8:19am

Thanks a lot this web is pretty good and the author writes in a very straight forward easy to understand manner.

I still dont quite get how to implement runge kutta though.

Ive seen how it works, and could implement it for one diferential equation, but in this case I have a serious doubt, since I have two ecuations (due to there being two unknowns) and on top of that the time is not directly in the actual ecuation, since the velocity and position are co dependant and time dependant as well I dont really know how to write it.

I've checked several pages:

http://myphysicslab.com/runge_kutta.html

and some others I cant recall, but I never see an example where the step size (in this case the time) is not explicitly in the ecuation.

Maybe its easier to tap into a dll that has a ODE solver but the only one I could get my hands on

http://www.alglib.net/

Doesnt have a Runge-Kutta solver, or at least I cant find it.

Thanks again for your help,

Permalink Reply by Nicolas on December 12, 2011 at 10:08am

Replying to my own discussion, haha, but maybe someone finds this useful.

I took some time off to do some other things, but now Im back on it, found an example of how to implement runge kutta to the spring ecuation, the guy is not the best speaker ever but it all makes sense now, I will try and implement it in VB.Net and if it works upload it here.

The final aim is doing a good wave simulation that can interact with objects and doesnt have to be an armonic function. I was starting to doubt this idea, thinking it might be stupid and I came across this video which is exactly what Im going for. Ill see if I can get it to work.

Permalink Reply by Jon Kniven on January 25, 2012 at 8:22am

I too could really use some wave simulation tools. How is the research going?

Permalink Reply by Nicolas on January 25, 2012 at 8:41am

I had to drop it, it was for a documentation and we ran out of time, when Im done with this I'll probably go back to it. I started using verlet solver to code springs, but i ran out of time pretty soon so I only got as far as linear spring system. It started giving me errors when I jumped to meshes, so I put it on hold. I dont know for when you need it, but i dont think Ill have anything running until a months time.

I upload the linear springs, dont know if it will me useful.

Attachments: