Method for processing special saddle point problem

Abstract

The invention discloses a method for processing a special saddle point problem, which converts an original linear system into an equivalent linear system easy to solve and enables a preprocessing matrix P-1A to be provided with better spectral properties than an original matrix, so that the convergence speed of the algorithm is increased, and the calculation efficiency of the algorithm is improved; and the shielding effect is improved, the external interference is reduced, and data errors are prevented.

Description

technical field [0001] The invention belongs to the technical field of saddle point problem processing, and more specifically relates to a processing method for a special saddle point problem. Background technique [0002] The saddle point problem is widely used in many scientific and engineering problems, such as fluid mechanics, optimization, hybrid finite element approximation of elliptic partial differential equations, etc. For the study of numerical solution methods for various types of saddle point problems, many scholars at home and abroad have given many effective numerical solution methods. When the scale of the problem is small, many classic methods can be effectively solved, such as LU decomposition method, Gauss-Seidel method, SOR method, conjugate gradient method, etc. [0003] However, when the order of the matrix is ​​large, many classical methods often cannot give satisfactory calculation results due to the influence of computer CPU calculation time, memory ...

AI Technical Summary

Problems solved by technology

[0004] However, with the rapid development of science and technology, the order of linear equations generated in practical problems is getting higher and higher. Using iterative methods to solve them will require a large amount of storage space, long computing time or super high-performance computers. There are many limitations in the application

At the same time, in the process of data processing, due to the influence of external interference, it will lead to output errors, which will affect the data results.

Images