Re: Jacobian with lsode

[Tatsuro MATSUOKA]

First I recommend to read the doument of  press
"lsode_options" from the prompt.
The the lsode solver uses the two different algorithms of
the ODE. One is for "non-stiff" ("Adams") and the other is
"stiff" ("BDF").
For the "non-stiff" case, the jacobian is not used.
For the "stiff" case, if the analytical jacobian is given,
the lsode use it but, if it is not given the lsode
calculate nummerical jacobian to solve the ODE.
Of course, the the "stiff" case, the analytical jacobian
is recommended.  But in many case, the automatically
computed nummerical jacobian works well.
I think that it is enough to give the analytical jacobian
when the lsode failes to solve the stiff ODE without the
analytical jacobian.

T. Matsuoka 

 


> 
> Hi all, 
> 
> nice forum. I am an Octave newbie but I guess my
> question is a bit silly. 
> I want to obtain the values for the jacobian at the
> solutions given by lsode
> for a certain function which I have already defined,
> let us say, as xdot =
> myfunction(x, t). 
> So, what I do is call x = lsode(["myfunction"; "j"],
> x0, t) having defined
> x0 and t. So, my question is : 
> do I have to define a function  jac = myj(x, t) and
> then run x =
> lsode(["myfunction"; "myj"], x0, t)
>  or does Octave compute the analytical/numerical
> jacobian itself? 
> Many thanks.
> 
> Alban
> -- 
> View this message in context:
>
http://www.nabble.com/Jacobian-with-lsode-tf4083078.html#a11604947
> Sent from the Octave - General mailing list archive
> at Nabble.com.
> 
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://www.cae.wisc.edu/mailman/listinfo/help-octave
> 


--------------------------------------
Easy + Joy + Powerful = Yahoo! Bookmarks x Toolbar
http://pr.mail.yahoo.co.jp/toolbar/