| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 6931247 | Journal of Computational Physics | 2015 | 12 Pages |
Abstract
By carefully exploring the structure of the coefficient matrix of the numerical method, we develop a preconditioned fast Krylov subspace method, which reduces the computations to (Nlogâ¡N) per iteration and the memory to O(N). The use of the preconditioner significantly reduces the number of iterations, and the preconditioner can be inverted in O(Nlogâ¡N) computations. Numerical results show the utility of the method.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Science Applications
Authors
Ning Du, Hong Wang, Che Wang,