I am solving a partial differential algebraic equation (PDAE) system which has the following dependent variables:

$f=f(X,T)$ and $g=g(T)$, along with a few others

My current method for coupling is to use fixed-point iteration with constant under-relaxation factors for the two coupled variables $f$ and $g$. The issue is that convergence (determined from the norm of the residuals) is too slow. To resolve this I have tried to implement dynamic under-relaxation using Aitken's method. However, this doesn't seem to offer any speed-up - it actually often requires more iterations to converge than using constant relaxation factors. This is the case even if I apply upper and lower bounds on the relaxation factors calculated by Aitken's method. Reducing the starting relaxation factors to small values (~0.01) also doesn't seem to help much.

My initial thoughts are that the Aitken's method is not accounting for the coupling between the two variables - the relaxations factors $\omega_f$ and $\omega_g$ are calculated with no consideration for the variation of each other. I have tried having a look at the literature, but I can only find the application of Aitken's method for dynamic relaxation for a single variable.

Any suggestions for improving the performance of Aitken's method in this case would be much appreciated


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.