A multipreconditioned conjugate gradient algorithm. (English) Zbl 1104.65027
The authors propose a generalization of the conjugate gradient method that uses multiple preconditioning, by combining them automatically in an optimal way. Their new algorithm obtains an energy norm minimization property, while maintaining \(A\)-conjugation and orthogonality properties similar to the preconditioned conjugate gradient method, but with iterates constructed in a generalized Krylov space incorporating an arbitrary set of preconditioners.
Reviewer: Constantin Popa (Constanţa)
MSC:
65F10 | Iterative numerical methods for linear systems |
65F35 | Numerical computation of matrix norms, conditioning, scaling |