Re: Solving unknowns of kinematic equations in octave.

[c.]

On 1 May 2013, at 17:45, ZachTM <address@hidden> wrote:

> Hello I am fairly new to octave but I have been playing around with fsolve. 
> I want to make an octave program that can solve for the 5 possible unknowns
> in the kinematic equations. (vi,vf,a,d, and t)
> 
> Would it be possible to somehow define these equations:
> <http://octave.1599824.n4.nabble.com/file/n4652529/KE1-3.jpg> 
> And then when I define 3 or more of the variables it could solve for the
> other two?  I just dont know how I would get this to work with fsolve.  I
> could make an octave program that picks which equation it will use for every
> unknown but I would like to extend this to work with any set of equations
> and unknowns (within reason of course).  If the user doesn't give enough
> variables I would return an error.
> 
> So can anyone tell me if this would be even possible?  And if fsolve would
> be appropriate to use here?  Any responses are greatly appreciated!
> 
> Thanks,
> Zach

It looks like the equations you want to solve are Newton's laws of motion?

That's a set of differentail equations, you may use an ode solver like LSODE, 
DASPK, DASSL 
or one of those included in the odepkg package in Octave-Forge.

For this specific problem the most appropriate class of solvers is that of 
symplectic integrators,
I have recently posted an initial implementation of one of those, the velocity 
Verlet method.

You should be able to find it by searching the mailing list archives.

HTH,
c.