Sparse Vector Methods

Abstract

Sparse vector methods enhance the efficiency of matrix solution algorithms by exploiting the sparsity of the independent vector and/or the desire to know only a subset of the unknown vector. This paper shows how these methods can be efficiently implemented for sparse matrices. The efficiency of the sparse vector methods is verified by tests on a 156-bus, a 1598-bus and a 2265-bus system. In all cases tested, the new methods are significantly faster than the established sparse matrix techniques.