
doi: 10.1137/040607964
This paper deals with a multilevel construction of hierarchical matrix approximations to the inverses of finite element stiffness matrices. Given a sequence of discretizations Aℓxℓ = f ℓ, ℓ = 0,..., n, where A0x0 = f 0 denotes the coarse grid problem, we will compute A0-1 exactly and then use interpolation to obtain an H-matrix approximation Aℓ+1-H from the approximate H-matrix inverse Aℓ-H on the next coarser grid. We develop an exact interpolation scheme for the inverse of tridiagonal matrices as they appear in the finite element discretization of one-dimensional differential equations. We then generalize this approach to two spatial dimensions where these efficiently computed approximations to the inverse may serve as preconditioners in iterative solution methods. We illustrate this approach with some numerical tests for convection-dominated convection-diffusion problems. © 2006 Society for Industrial and Applied Mathematics.
Inverses of tridiagonal matrices, Hierarchical matrices, Data-sparse approximation, Multilevel
Inverses of tridiagonal matrices, Hierarchical matrices, Data-sparse approximation, Multilevel
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 2 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Average | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
