By Robert M. Corless

This ebook offers an intensive creation to numerical computing from the perspective of backward errors research. The meant viewers comprises scholars and researchers in technology, engineering and arithmetic. The technique taken is a bit casual because of the big variety of backgrounds of the readers, however the primary rules of backward mistakes and sensitivity (conditioning) are systematically emphasised. The publication is split into 4 elements: half I presents the historical past preliminaries together with floating-point mathematics, polynomials and machine overview of services; half II covers numerical linear algebra; half III covers interpolation, the FFT and quadrature; and half IV covers numerical ideas of differential equations together with initial-value difficulties, boundary-value difficulties, hold up differential equations and a quick bankruptcy on partial differential equations.

The publication includes special illustrations, bankruptcy summaries and quite a few workouts besides a few Matlab codes supplied on-line as supplementary material.

“I quite just like the specialize in backward errors research and situation. this is often novel in a textbook and a pragmatic method that may carry welcome attention." Lawrence F. Shampine

A Graduate creation to Numerical equipment and Backward blunders research” has been chosen by way of Computing experiences as a amazing ebook in computing in 2013. Computing stories better of 2013 record contains e-book and article nominations from reviewers, CR classification editors, the editors-in-chief of journals, and others within the computing community.

Show description

Read or Download A Graduate Introduction to Numerical Methods: From the Viewpoint of Backward Error Analysis PDF

Similar introduction books

Color: An Introduction to Practice and Principles, Second Edition

Content material: bankruptcy 1 resources of colour (pages 1–16): bankruptcy 2 what's colour and the way Did We Come to event It? (pages 17–25): bankruptcy three From mild to paint (pages 26–36): bankruptcy four colour belief: Phenomena (pages 37–52): bankruptcy five Orderly preparations of colour (pages 53–77): bankruptcy 6 Defining the colour Stimulus (pages 78–92): bankruptcy 7 Calculating colour (pages 93–105): bankruptcy eight Colorants and Their blend (pages 106–118): bankruptcy nine colour replica (pages 119–129): bankruptcy 10 the net of colour (pages 130–147): bankruptcy eleven colour (Theory) in paintings (pages 148–160): bankruptcy 12 concord of colours (pages 161–170):

Introduction to Topology and Geometry, Second Edition

An simply available advent to over 3 centuries of thoughts in geometryPraise for the 1st Edition“. . . a welcome replacement to compartmentalized remedies certain to the outdated pondering. This sincerely written, well-illustrated booklet provides adequate historical past to be self-contained. ” ? CHOICEThis absolutely revised re-creation bargains the main complete insurance of contemporary geometry presently on hand at an introductory point.

Getting Started in Bonds, Second Edition

Every thing you must find out about bondsBonds are a key part in each portfolio, making it crucial that traders comprehend what precisely they're and the way they functionality. Getting begun in Bonds, moment version has been totally up to date and revised take into consideration the ever-changing bond marketplace in addition to the present financial atmosphere.

Extra info for A Graduate Introduction to Numerical Methods: From the Viewpoint of Backward Error Analysis

Sample text

23) n=0 (see Bender and Orszag (1978) and Chap. 3 of this book). , until adding new terms does not change the solution anymore because they are too small). Of course, true convergence would require that, for every ε > 0, there existed an N such that ∑M k≥N+1 ak < ε for any M > N, that is, that the sequence of partial sums was a Cauchy sequence. There are many tests for convergence. Indeed, for this Taylor series, we can easily use the Lagrange form of the remainder and an accurate plot of the 31st derivative of the Airy function on this interval to establish that 30 terms in the series has an error less than 10−16 on the interval −12 ≤ z ≤ 4.

282 Discrete Laplacian on 20 × 20 C grid . . . . . . . . . . . . . . . 283 The implicit curves 25xy = 12 and x2 + y2 = 1 . . . . . . . . . . . 3 Scaled residuals for the Richardson iteration of a 5 × 5 matrix . . . . 314 Scaled residuals for the Richardson iteration of a 89 × 89 matrix . . . 315 Nine hundred ninety-nine eigenvalues of a random 1, 000 × 1, 000 real matrix . . . . . . . . . . . . . . . .

201 Residual in SVD computed by MATLAB’s effective algorithm . . . . 203 A close-up of a tawny frogmouth . . . . . . . . . . . . . . . . 209 The tawny frogmouth, with only 25 nonzero singular values . . . . . 210 Computing time for PA = LU . . . . . . . . . . . . . . . . . . 217 Scaled residual in PA − LU / A for some matrices . . . . . . . . . . 218 Average timings for various factorings . . . . . . . . . . . . . . 2 Initial growth of matrix powers of a 50 × 50 matrix .

Download PDF sample

Rated 4.31 of 5 – based on 29 votes