© 1968 by British Computer Society
| ||||||||||||||||||||||||||||||||||||||||||||||||||
An empirical estimate of the relative error of the computed solution x of Ax = b
Computer Center, Queen Mary College, London, UK
An empirical method is described for providing, with a minimum of effort, useful estimates for the relative errors in individual components of the computed solution x of the system of linear equations Ax = b. The estimate for the error is in no sense a rigorous upper bound, but in view of the economy of effort it is felt that extensions of the method described here a valuable addition to known methods of a posteriori error analyses.
First received June 1967. revised form, September 1967.
* Computer Center, Queen Mary College, London, E.1.
Present address: Mathematics Department, Birkbeck College, London, W.C.1.