Re: lsode for matrix DGLs?

[Matthias Brennwald]

On 22.06.2007, at 19:07, address@hidden wrote:

> I think you should use the ode's functions for that.

What do you mean by 'that'?

> the lsode calculate
> the jacob matrix, and you need partials, not only dt,
> I hope this example will help.
>
> Lotka- Volterra equations:
> dx/dt=x*(a-b*y)
> dy/dt=-y*(c-d*x)
>
> making a function:
>
> function ydot = lotkavolterra(t,y)
> global a b c d
> A= [ a                 -b*y(1)
>          d*y(2)         -c         ];
> ydot = A*y;
>
> in octave:
> octave>a=1.5;b=1;c=3;d=1;
> [t,y]=ode23(@lotkavolterra,[0 20],[2 2]);

Uhm, ode23 is from the competition, right?

> plot(t,y(:,1))
> hold on
> plot(t,y(:,2))

So, y1 = y(:,1) and y2 = y(:,2) are two vectors reflecting two  
_scalar_ functions of time t: y1(t) and y2(t)
What I'm after is two (or more) _matrix_ functions of time. I guess  
Johns approach of reshaping my matrices to vectors should do the trick.

Cheers
Matthias




-------
Matthias Brennwald
Lägernstrasse 6
CH 8037 Zürich
+41 (0)44 364 17 03
address@hidden