Mode ordering algorithms for sparse vector method improvement

Three node-ordering algorithms that enhance sparse vector methods without losing sparse matrix method efficiency are presented. The performance of the algorithms is compared with that of the minimum-degree algorithm for eight test systems up to 661 nodes. For the factorization of a sparse matrix into U/sup t/U, the new algorithms give a similar or even better sparsity for U, while clearly improving the sparsity of U/sup -1/.<
>