Abstract
Kron's techniques of tearing and k-partitioning are discussed as they apply to large linear systems of an arbitrary nature and tearing is discarded as inefficient. It is shown that decreasing the size of the sub-units in k-partitioning increases its efficiency; carried to the limit, this reduction results in a Gaussian elimination scheme. Finally, the optimum application of k-partitioning is presented as a linear programming problem.

This publication has 3 references indexed in Scilit: