[SciPy-user] integrate ODE to steady-state?

Zachary Pincus zachary.pincus@yale....
Tue Jun 24 18:45:10 CDT 2008


Hi all,

So, after a brief bout of the stupids (thanks Robert), I have  
formulated my optimization problem as a physical system governed by an  
ODE, and I wish to learn the equilibrium configuration of the system.

Any thoughts on what the easiest way to do this with scipy.integrate  
is? Ideally, I'd just want the solver to take as large steps as  
possible until things converge, and so I don't really care about the  
"time" values. One option would be to use odeint and just tell it to  
integrate to a distant time-point when I'm sure things will be in  
equilibrium, but that seems dorky, wasteful, and potentially incorrect.

Alternately, I could use the ode class and keep asking it to integrate  
small time-steps until the RMS change drops below a threshold. There,  
still, I'd need to choose a reasonable time-step, and also the inner  
loop would be in python instead of fortran.

Any recommendations? (Or a I again being daft? I never really took a  
class in numerical methods, so sorry for dim-bulb questions!)

Zach


More information about the SciPy-user mailing list