Accurate computation of the Moore-Penrose inverse of strictly totally positive matrices

The computation of the Moore-Penrose inverse of structured strictly totally positive matrices is addressed. Since these matrices are usually very ill-conditioned, standard algorithms fail to provide accurate results. An algorithm based on the factorization and which takes advantage of the special st...

ver descrição completa

Detalhes bibliográficos
Autores: Marco García, Ana|||0000-0003-4662-6327, Martínez Fernández De Las Heras, José Javier|||0000-0002-8322-0361
Formato: artículo
Fecha de publicación:2019
País:España
Recursos:Universidad de Alcalá (UAH)
Repositorio:e_Buah Biblioteca Digital Universidad de Alcalá
Idioma:inglés
OAI Identifier:oai:ebuah.uah.es:10017/60638
Acesso em linha:http://hdl.handle.net/10017/60638
https://dx.doi.org/10.1016/j.cam.2018.10.009
Access Level:acceso abierto
Palavra-chave:Moore Penrose inverse
Inverse
Totally positive matrix
Neville elimination
Bidiagonal decomposition
High relative accuracy
Descrição
Resumo:The computation of the Moore-Penrose inverse of structured strictly totally positive matrices is addressed. Since these matrices are usually very ill-conditioned, standard algorithms fail to provide accurate results. An algorithm based on the factorization and which takes advantage of the special structure and the totally positive character of these matrices is presented. The first stage of the algorithm consists of the accurate computation of the bidiagonal decomposition of the matrix. Numerical experiments illustrating the good behavior of our approach are included.Numerical experiments illustrating the good behavior of our approach are included.