Approximate Matrix Inversion by Aggregation
The so called aggregation method has been suggested in  for inverting block matrices which may be split into the sum of two matrices, one of them being invertible and the other one being “Block-wise constant”, i,e. having in each of its blocks all elements of the same value.The method consists in replacing the inversion of a large block matrix by the inversion of a matrix of a lower dimension. It was studied in [ 3] from the numerical point of view and several numerical examples have been performed. Some results concerning the applicability of this method — in particular to certain class of sparse Leontjev’s matrices — are presented in this paper.
Unable to display preview. Download preview PDF.
- 1.Fiedler, M. -Pták V. : On Aggregation in Matrix Theory and Its Application to Numerical Inverting of Large Matrices, Bulletin dell’acad. Polon. des Sc., Série des sc. math., astr. et phys., vol. XI, No 12 (1963) 757 -759Google Scholar
- 2.Nováková, M. : On the Use Of Some Properties Of Leontjev’s Matrices In Their Inversion, Aplikace matematiky, Svazek 15,. 2, (1970)Google Scholar
- 3.Prágerová, I. : Poušití agpegace p;i inversi matic, Praha (1980), (Diploma work)Google Scholar