Customizable triangular factorizations of matrices
β Scribed by Pengwei Hao
- Publisher
- Elsevier Science
- Year
- 2004
- Tongue
- English
- Weight
- 229 KB
- Volume
- 382
- Category
- Article
- ISSN
- 0024-3795
No coin nor oath required. For personal study only.
β¦ Synopsis
Customizable triangular factorizations of matrices find their applications in computer graphics and lossless transform coding. In this paper, we prove that any N Γ N nonsingular matrix A can be factorized into 3 triangular matrices, A = PLUS, where P is a permutation matrix, L is a unit lower triangular matrix, U is an upper triangular matrix of which the diagonal entries are customizable and can be given by all means as long as its determinant is equal to that of A up to a possible sign adjustment, and S is a unit lower triangular matrix of which all but N -1 off-diagonal elements are set zeros and the positions of those N -1 elements are also flexibly customizable, such as a single-row, a single-column, a bidiagonal matrix or other specially patterned matrices. A pseudo-permutation matrix, which is a simple unit upper triangular matrix with off-diagonal elements being 0, 1 or -1, can take the role of the permutation matrix P as well. In some cases, P may be the identity matrix. Besides PLUS, a customizable factorization also has other alternatives, LUSP, PSUL or SULP for lower S, and PULS, ULSP, PSLU, SLUP for upper S.
π SIMILAR VOLUMES
In this paper we give explicit factorizations which demonstrate the stable tameness of all polynomial automorphisms arising from a recent construction of Hubbers and van den Essen. This is accomplished by two different factorizations of such an automorphism by triangular automorphisms, one which is
This Paper analyzes the factorization of the inverse of a Cauchy-Vandermonde matrix as a product of bidiagonal matrices. Factorizations of Cauchy-Vandermonde matrices are also considered. The results are applied to derive a fast algorithm for solving linear Systems whose coefficient matrices are Cau