Title :
Generalized inverses of large matrices using the generalized Schur complement
Author :
Frank, Neil ; Imam, Lbrahim
Author_Institution :
Dept. of Eng. Math. & Comput. Sci., Louisville Univ., KY, USA
Abstract :
The development of a software package for calculating the generalized inverse of a large real matrix is described. A large matrix is defined as a matrix that is too large to reside in computer memory. The computer software is written in Microsoft C v5.0 and can be implemented on an IBM-PC or compatible. The software is file-oriented and employs windowing techniques developed to bring into RAM the block needed for the calculations that are performed. The program requires that all matrices be stored in disk files before any manipulation can take place. This design constraint is necessary because when performing calculations on large matrices, only part of the matrix can be read into memory at any given time. Details about the design structure of the software are given. An additional set of functions is developed to allow the user to use this program in a recursive fashion in order to calculate the M-P inverse of extremely large matrices
Keywords :
IBM computers; file organisation; mathematics computing; matrix algebra; microcomputer applications; IBM-PC; M-P inverse; Microsoft C; RAM; computer software; design constraint; design structure; disk files; file-oriented; generalized Schur complement; generalized inverse; large real matrix; recursive fashion; software package; windowing techniques; Computer science; Equations; Mathematics; Matrix decomposition; Null space; Read-write memory; Singular value decomposition; Software design; Software packages;
Conference_Titel :
Southeastcon '90. Proceedings., IEEE
Conference_Location :
New Orleans, LA
DOI :
10.1109/SECON.1990.117768
