Re: Ode45 help

[Thomas Treichl]

Brandon Pye schrieb:
> On Saturday 14 February 2009 16:51:15 you wrote:
> > Brandon Pye schrieb:
> > > Hey guys,
> > >
> > > I'm in a class where I need to use ode45. I keep getting an error
> > > syntax when I copy the code from matlab to octave. The function that
> > > contains the differential equations gets its values from a structure
> > > defined in the master .m file. In matlab, when I call ode45 the code
> > > looks like:
> > >
> > > [t,y] = ode45(@gene.m, [0 10], y0, [], info);
> > >
> > > Where the closed brackets are skipping the tolerance and 'info' is the
> > > structure I need so that the function 'gene.m' can get the values
> > > necessary for the differential equations. When I run that code in
> > > octave, it can't find the values in the structure. Can anyone tell me
> > > where I can put the structure in the code when I call ode45 in octave?
> > > Also, I have matlab for windows. Should I just try to run it with wine?
> > I took a similar example than the one you sent (a modified version of the
> > example that comes from "help ode45" at the Octave prompt. I don't know
> > what the gene function should do but I also cannot see the problem. Can you
> > post some more information from your session?
> >
> > Here is my example:
> >
> >    octave-3.0.3:1> fvdb = @(vt,vy) [vy(2); (1 - vy(1)^2) * vy(2) - vy(1)];
> >    octave-3.0.3:2> ode45 (fvdb, [0 20], [2 0], [], (struct ('a', 1, 'b',
> > 2)))
> >
> > Best regards,
> >
> >    Thomas
>
> The master equation uses the code:
> [t,y] = ode45(@gene_iffl1, [0 info.tmax], info.y0, [], info);
>
> with info containing:
> info.gen_time = 0.5
> info.alpha_y = log(2)/info.gen_time;
> info.alpha_z = log(2)/info.gen_time;
> info.beta_zhigh = 50;
> info.beta_zlow = 2;
> info.beta_y = 50;
> info.K_yz = 20;
> info.tmax = 10;
>
> The function gene_iffl1 is:
>
> function dydt = gene_iffl1(t,y,info)
> dydt = zeros(2,1);
> dydt(1) = info.beta_y-info.alpha_y*y(1);
> dydt(2)=info.beta_zhigh*(1-(info.beta_zhigh-
> info.beta_zlow)/info.beta_zhigh*(y(1)>info.K_yz))-info.alpha_z*y(2);
>
> I just need help with translating the syntax essentially.

Everything you do is ok. The only thing that need to be done is that you pass an 
empty pseudo options handling structure (which is generated with the command 
'odeset') as the 4th input argument. In your case, the solver somehow doesn't 
like an empty 4th input argument if the 5th input argument is a structure, too. 
I must have a look at the solver function some when I have the time to do it.

Your workaround that works in Matlab and Octave should be

  vopt = odeset ();
  [t,y] = ode45(@gene_iffl1, [0 info.tmax], info.y0, vopt, info);

BTW, take care with your info variable which also is a valid Octave function!

Best regards,

  Thomas