[SciPy-User] Why does scipy.integrate.odeint still provide a result even after "repeated convergence failures"?

[Sturla Molden]

Are you trying to fit an ODE or just solve an ODE with known parameters?

Fitting parameters by solving an ODE can be a notoriously difficult 
problem with respect to convergence. There are methods such as "multiple 
shooting" which can improve the situation. But fitting an ODE is often 
an ill-posed problem.


Sturla



On 12/05/15 07:44, Brian Merchant wrote:
> Hi Warren,
>
> Just an update: I was able to "zero" in on the portion of the model that
> causes the issue (I set the ode function to just return 0 for every
> other variable, except the one that causes issues).
>
> I can't find an error in the Python code responsible for calculating the
> derivative for that variable (I stepped through it, it simply implements
> a Hill equation:
> http://en.wikipedia.org/wiki/Hill_equation_%28biochemistry%29), and the
> results provided by odeint in the end don't  seem *obviously* out-of-whack.
>
> Brian
>
> On Mon, May 11, 2015 at 9:52 PM, Warren Weckesser
> <warren.weckesser at gmail.com <mailto:warren.weckesser at gmail.com>> wrote:
>
>
>
>     On Tue, May 12, 2015 at 12:43 AM, Brian Merchant
>     <bhmerchant at gmail.com <mailto:bhmerchant at gmail.com>> wrote:
>
>         Hi all,
>
>         Why does scipy.integrate.odeint still provide a result even
>         after "repeated convergence failures"? Is this result still of use?
>
>
>
>     Which version of scipy are you using?  Can you give a short example
>     that demonstrates the problem?
>
>     Warren
>
>
>         Kind regards,
>         Brian
>
>         _______________________________________________
>         SciPy-User mailing list
>         SciPy-User at scipy.org <mailto:SciPy-User at scipy.org>
>         http://mail.scipy.org/mailman/listinfo/scipy-user
>
>
>
>     _______________________________________________
>     SciPy-User mailing list
>     SciPy-User at scipy.org <mailto:SciPy-User at scipy.org>
>     http://mail.scipy.org/mailman/listinfo/scipy-user
>
>
>
>
> _______________________________________________
> SciPy-User mailing list
> SciPy-User at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>