-
-
Notifications
You must be signed in to change notification settings - Fork 5.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Radau - missing option "mass" #12762
Comments
A1: No, mass matrix option is not implemented. A2: If someone takes the initiative to implement that. When we considered implicit ODEs in Now as |
@nmayorov Many thanks for your reply. |
The system closed this issue and attributed the closure to me - some how I mistakenly hit the wrong button :-(. |
A Matlab version of RadIIA is also available at: https://uk.mathworks.com/matlabcentral/fileexchange/56162-radau-iia |
Hi, |
Hi,
Thanks for the update regarding Radau and the mass matrix addition.
Glad that you are making progress. Please continue to keep me informed.
Kind regards,
Graham W Griffiths
|
Hi, |
Hi, here's an update. I've made a Git with my modifications to Scipy's Radau and a few test cases: I have added various printouts in the code to analyse the evolution of the Newton solution, the error estimate and other parameters. Various test problems are given, which attest that the new version of Radau is able to solve DAE up to index 3. These tests include:
Maybe I'll add a discretised PDE next. I'll also verify the order of convergence on some of the test cases. Following my read of "The numerical solution of differential-algebraic systems by Runge-Kutta methods" by Hairer, Lubich, Roche, I have decided to force to 0 the error estimate for the algebraic variables. These variables are at the moment identified as those that correspond to nil rows in the mass matrix. The transistor amplifier does not allow for such an identification, however it works fine without this trick. The authors of the previously mentioned book recommend multiplying the errors on the algebraic variables by the time step h (index-1), or h^2 (index 2). Is the Scipy team willing to integrate this modification of the solver ? :) Regards |
Ability to solve ODEs or DAEs with a mass matrix.
I am trying to use scipy.integrate.solve_ivp for a system with a mass matrix 'M', but setting method='Radau' and using **options as options={'mass':M} throws the warning:
UserWarning: The following arguments have no effect for a chosen solver:
mass
.Q1: is the mass matrix option implemented in Radau?
Q2: If the mass matrix option is not implemented, is it likely to be added to Radau or any other solver in the near future?
Thanks for considering these questions.
The text was updated successfully, but these errors were encountered: