analysis of pivot strategies to maintain sparsity in the LU decomposition of IPDG method applied to the Helmholtz Equation
Ryan Samuel Severance
University of West Florida,
Master of Science (MS), University of West Florida
2019
Metrics
3 File views/ downloads
90 Record Views
Abstract
In recent years, the interior penalty discontinuous Galerkin (IPDG) method has appeared in literature as an efficient and stable method for approximating the Helmholtz equation. LU decomposition has then been used to solve the linear system formed by the IPDG method. However, research has shown that the LU decomposition causes fill-in of the sparse structure of the global matrix. This talk addresses the application of several pivot strategies to the global matrix before the LU decomposition, in order to assess if this fill-in can be reduced. Numerical experiments are presented to demonstrate that pivot strategies did reduce fill-in when applying the LU decomposition.