Perturbation Methods for the Solution of Linear Problems.
Abstract
Linear problems of central interest in numerical analysis are the solution of linear equations, the construction of the inverse or a generalized inverse of a linear operator, finding the eigenvalues and eigenvectors of a linear operator, and linear programming. A survey is made of methods which apply if the data of a solved linear problem is perturbed by operators and vectors of small norm (analytic perturbation), or by operators of finite rank and vectors belonging to a finite-dimensional subspace (algebraic perturbation). Perturbation methods may be used to extend the theory of linear problems, to obtain economy of effort in the solution of perturbed problems, and to estimate errors due to inaccurate data and computation. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1977
- Accession Number
- ADA044829
Entities
People
- Louis B. Rall
Organizations
- University of Wisconsin–Madison