|
Such methods are based upon in the krylov subspace $`\Kappa =span\{ r_0,..., A^{(k-1)}r_0\}`$, with $`r_0`$ being the residual $`r_0=b-Ax_0`$ for the $`k=0`$ iteration. Particularly GMRES minimizes $`||r_k||_2`$ in $`x_0+\Kappa(A,r_0)`$.
|
|
Such methods are based upon in the krylov subspace $`\Kappa =span\{ r_0,..., A^{(k-1)}r_0\}`$, with $`r_0`$ being the residual $`r_0=b-Ax_0`$ for the $`k=0`$ iteration. Particularly GMRES minimizes $`||r_k||_2`$ in $`x_0+\Kappa(A,r_0)`$.
|