6
$\begingroup$

Let's say you have a general matrix $A$, with diagonal entries $a_{ii} = d>0$. (No assumptions are made about the off-diagonal elements.) Then Jacobi preconditioning doesn't improve condition number, because $$ \kappa(A)= \left|\frac{\lambda_{\rm{max}}(A)}{\lambda_{\rm{min}}(A)} \right| $$ becomes $$ \kappa(M^{-1}A)= \left|\frac{\lambda_{\rm{max}}(A)/d}{\lambda_{\rm{min}}(A)/d} \right| = \kappa(A). $$ In this case, could the use of a Jacobi preconditioner do anything useful?

$\endgroup$
6
$\begingroup$

No, Jacobi only ever corrects relative scales. It does nothing for "smooth" ill-conditioning, such as the $\kappa(A) \in O(h^{-2})$ asymptotics for second order elliptic problems.

If you are using a Krylov method, the global scale is automatically corrected, but with a stationary iteration, the (constant) scaling is needed somehow (could just be in the damping parameter for the stationary method).

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.