Iterative Methods for Large Sparse Linear Systems Arising from Partial Differential Equations
Abstract
The objective of this project was to study numerical methods for solving sparse linear systems of equations of the type that arise from discretized partial differential equations. Such systems arise in mathematical models of numerous physical processes including turbulent flow, chemical reactive flow, semiconductor device simulation, and structural mechanics. For many such problems, analytic solutions are not available, so the only way of obtaining insight into the model is through numerical approximation and solution. The solution of sparse linear systems is often the most costly task, in terms of both computer time and storage, required by this process, so that improving the efficiency of such computations is of critical importance for the construction of accurate numerical models. Emphasis has been on sparse iterative methods for solving these systems, and implementation of linear system solvers on parallel computers. Major results are summarized in the final report.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1992
- Accession Number
- ADA249524
Entities
People
- Howard C. Elman
Organizations
- University of Maryland