An approximate minimum degree algorithm for matrices with dense rows
نویسندگان
چکیده
We present a modified version of the approximate minimum degree algorithm for preordering a matrix with a symmetric sparsity pattern prior to the numerical factorization. The modification is designed to improve the efficiency of the algorithm when some of the rows and columns have significantly more entries than the average for the matrix. Numerical results are presented for problems arising from practical applications and comparisons are made with other implementations of variants of the minimum degree algorithm.
منابع مشابه
A note on fast approximate minimum degree orderings for symmetric matrices with some dense rows
Recently a number of variants of the approximate minimum degree algorithm have been proposed that aim to efficiently order symmetric matrices containing some dense rows. We compare the performance of these variants on a range of problems and highlight their limitations. This leads us to propose a new variant that offers both speed and robustness.
متن کاملAlgorithm 8xx: AMD, an approximate minimum degree ordering algorithm
AMD is a set of routines for permuting sparse matrices prior to numerical factorization, using the approximate minimum degree ordering algorithm. There are versions written in both C and Fortran 77. A MATLAB interface is included.
متن کاملToward an Efficient Column Minimum Degree Code for Symmetric Multiprocessors
Ordering the columns of a nonsymmetric sparse matrix can reduce the fill created in its factorization. Minimum-degree is a popular heuristic for ordering symmetric matrices; a variant that can be used to order nonsymmetric matrices is called column minimum degree. In this paper we describe the design of a multithreaded approximate column minimum degree code. We present a framework for the algor...
متن کاملImproving the Run Time and Quality of Nested Dissection Ordering
When performing sparse matrix factorization, the ordering of matrix rows and columns has a dramatic impact on the factorization time. This paper describes an approach to the reordering problem that produces significantly better orderings than prior methods. The algorithm is a hybrid of nested dissection and minimum degree ordering, and combines an assortment of different algorithmic advances. N...
متن کاملProbabilistic Matrix Addition
We introduce Probabilistic Matrix Addition (PMA) for modeling real-valued data matrices by simultaneously capturing covariance structure among rows and among columns. PMA additively combines two latent matrices drawn from two Gaussian Processes respectively over rows and columns. The resulting joint distribution over the observed matrix does not factorize over entries, rows, or columns, and can...
متن کامل