scientific article; zbMATH DE number 11411

From MaRDI portal
Publication:3974342





The solution of a linear system of equations often requires the computations of the generalized inverse of a matrix. The process is very time consuming. An efficient computational method is of particular interest when we have a large scale system. This paper presents an efficient computation method for solving a linear system with arrow block structure. This class of problems occurs quite frequently in the studies of planning and mechanics. In order to avoid the need to compute the generalized inverse, this paper employs the deflation technique to single out the minimal singular value and its associated singular vector for each submatrix in the diagonal block. These left and right singular vectors are then used to facilitate the solution in an algorithm which is computationally efficient on parallel processing in multiprocessors. Numerical examples are also given to illustrate its applications. Only solution errors are shown in the examples. The paper would be more interesting if comments are given on computational time, especially when it is compared with the other methods that use the generalized inverse.












This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3974342)