Nicholas J. Higham's Accuracy and Stability of Numerical Algorithms, Second PDF

By Nicholas J. Higham

ISBN-10: 0898715210

ISBN-13: 9780898715217

Show description

Read Online or Download Accuracy and Stability of Numerical Algorithms, Second Edition PDF

Best computational mathematicsematics books

Get Computational and Numerical Challenges in Environmental PDF

Many huge mathematical types, not just types bobbing up and utilized in environmental reports, are defined by way of structures of partial differential equations. The discretization of the spatial derivatives in such versions ends up in the answer of very huge platforms of normal differential equations. those structures include many hundreds of thousands of equations and feature to be dealt with over huge time periods through using many time-steps (up to a number of hundred thousand time-steps).

Download e-book for kindle: Logic and Computational Complexity: International Workshop by Felice Cardone (auth.), Daniel Leivant (eds.)

This booklet comprises revised types of papers invited for presentation on the overseas Workshop on good judgment and Computational Complexity, LCC '94, held in Indianapolis, IN in October 1994. The synergy among common sense and computational complexity has won significance and vigour in recent times, slicing throughout many parts.

Extra resources for Accuracy and Stability of Numerical Algorithms, Second Edition

Example text

I=l Computing s~ from this formula requires two passes through the data, one to compute x and the other to accumulate the sum of squares. A two-pass computation is undesirable for large data sets or when the sample variance is to be computed as the data is generated. 5) This formula is very poor in the presence of rounding errors because it computes the sample variance as the difference of two positive numbers, and therefore can suffer severe cancellation that leaves the computed answer dominated by roundoff.

Note that this definition is specific to problems where rounding errors are the dominant form of errors. The term stability has different meanings in other areas of numerical analysis. 2. Mixed forward-backward error for y = computed. = f(x). 6. Conditioning The relationship between forward and backward error for a problem is governed by the conditioning of the problem, that is, the sensitivity of the solution to perturbations in the data. :1x). :1X)2, 2. OE(O,I), and we can bound or estimate the right-hand side.

For illustration, consider the system Ax = b, where A is the inverse of the 5 x 5 Hilbert matrix and bi = (-l)ii. ) We solved the system in varying precisions with unit roundoffs u = 2- t , t = 15: 40, corresponding to about 4 to 12 decimal places of accuracy. ) The algorithm used was Gaussian elimination (without pivoting), which is perfectly stable for this symmetric positive definite matrix. 3 shows t against the relative errors Ilx - xlloo/llxll oo and the relative residuals lib - Axlloo/(IIAlloo Ilxll oo ).

Download PDF sample

Accuracy and Stability of Numerical Algorithms, Second Edition by Nicholas J. Higham

by William

Rated 4.04 of 5 – based on 34 votes