Estabilidade dos algoritmos da álxebra lineal numérica
Loading...
Identifiers
Publication date
Authors
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A análise do erro cara atrás, técnica principalmente desenvolta e popularizada por
James H. Wilkinson, baséase en explicar os erros de redondeo na solución calculada dun
problema matemático como consecuencia dunha lixeira perturbación dos respectivos datos
de entrada. No Capítulo 1 incluímos algunhas definicións, notacións e resultados acerca
da aritmética en coma flotante, nos que se fundamenta o resto do traballo. No Capítulo 2
describimos como se leva a cabo a análise do erro cara atrás e relatamos un par de exemplos
históricos nos que o seu uso foi esencial. Tamén introducimos algúns conceptos e resultados
útiles para interpretala, como a estabilidade cara atrás. Nos Capítulos 3 e 4 dánse algúns
exemplos de algoritmos estables cara atrás aplicados a dous problemas fundamentais da
Álxebra Lineal Numérica: a resolución de sistemas de ecuacións lineais e o cálculo de valores
propios.
Backward error analysis, a method mainly developed and popularized by James H. Wilkinson, consists on explaining rounding errors in the computed solution of a mathematical problem as a consecuence of a small perturbation on the corresponding input data. In the first chapter, we include some definitions, notations and results about floating point arithmetic that will be the basis for the rest of this paper. In Chapter 2 we describe how backward error analysis is carried out and we give some historical examples where its use was essential. We also introduce some useful concepts and results for its interpretation, such as backwards stability. In Chapter 3 and 4 we present some examples of backward stable algorithms applied to two fundamental problems of Linear Numerical Algebra: the resolution of linear systems and eigenvalue problems.
Backward error analysis, a method mainly developed and popularized by James H. Wilkinson, consists on explaining rounding errors in the computed solution of a mathematical problem as a consecuence of a small perturbation on the corresponding input data. In the first chapter, we include some definitions, notations and results about floating point arithmetic that will be the basis for the rest of this paper. In Chapter 2 we describe how backward error analysis is carried out and we give some historical examples where its use was essential. We also introduce some useful concepts and results for its interpretation, such as backwards stability. In Chapter 3 and 4 we present some examples of backward stable algorithms applied to two fundamental problems of Linear Numerical Algebra: the resolution of linear systems and eigenvalue problems.
Description
Traballo Fin de Grao en Matemáticas. Curso 2021-2022
Keywords
Bibliographic citation
Relation
Has part
Has version
Is based on
Is part of
Is referenced by
Is version of
Requires
Sponsors
Rights
Atribución-NoComercial-CompartirIgual 4.0 Internacional



