Multi-wavefront block ilu preprocessing method based on inverse

Abstract

The invention discloses an inverse-based multi-wave front block ILU preprocessing method, which belongs to the field of numerical solution, and aims to overcome the problem of low efficiency of traditional ILU preprocessing when solving large sparse linear equations. The present invention mainly adopts the discarding strategy based on the inverse, which greatly reduces the occurrence of numerical instability caused by discarding elements, and improves the numerical stability; adopts the multi-wavefront method that does not need to store the update matrix, avoids The temporary storage update array greatly reduces the memory overhead; the superblock-adaptive block incomplete decomposition method is adopted, which greatly improves the computing performance and memory performance. Therefore, the inverse-based multi-wavefront block ILU preprocessing method proposed by the present invention can greatly improve the efficiency of solving large-scale sparse asymmetric linear equations, and obtain large-scale sparse asymmetric linear equations more accurately and quickly with a small memory cost. Solutions to systems of linear equations.

Description

technical field [0001] The invention belongs to the field of numerical solution, and in particular relates to an inverse-based multi-wave front block ILU preprocessing method, which is used for solving large-scale sparse linear equations generated in scientific and engineering calculation problems. Background technique [0002] These large-scale , Very large scientific and engineering problems are very complex. At this time, traditional experimental methods and trial methods are not only costly, but also take a lot of time; because it is difficult to give the functional relationship between various physical quantities in these problems, it is impossible to obtain accurate solutions to these complex problems through analytical methods. Compared with the analytical method, the numerical method does not solve the problem through the specific mathematical function relationship, but approximates the solution of the problem through a series of specific values ​​distributed on sev...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the problem of low efficiency when traditional ILU preprocessing solves large sparse linear equations, and proposes an inverse-based multi-wave front block ILU preprocessing method, which can be more accurate and efficient at a smaller memory cost. Efficiently find solutions to large sparse systems of linear equations

Images