On Optimal-Pivoting Algorithms in Sparse Matrices

Abstract
This correspondence presents an alternate method for finding the minimum fill-in during each step of elimination for both the GaussJordan elimination and the Gaussian elimination. The correspondence also describes an improvement in sparsity in computer-aided large-network analysis by obtaining the optimal pivot at each step of elimination on sparse matrices. This proposed optimal-pivot ordering appears to lead to an overall minimum for computer storage and arithmetic operation count.