Numerical Methods for Scientists and Engineers, Third Edition

During the last twenty years since the publication of the first edition of the book, the speed as well as memory of computers have increased by a few orders of magnitude. However, the precision with which the floating point operations are being handled on these computers has not improved. As a result, the relevance of roundoff errors in numerical computations has increased substantially. In this book, an attempt is made to demonstrate that with proper care the errors do not grow very fast with the size of computational problem. Thus the book is probably more relevant today than it was twenty years back. The main aim of this book has been to not only provide suitable algorithms for numerical computations, but also to explain their limitations.
Based on experience in teaching, some new topics have been added to the second edition. A notable omission in the second edition of the book was statistical analysis of data. This has been addressed by the addition of a new Chapter (Chapter 9) on Statistical Inference. This Chapter introduces basic statistical distributions, which are useful in the next Chapter on approxima- tions. In view of this addition, the initial part of Chapter 10 on approximation has also been edited and two new subsections on least squares fit when there are errors in both x and y, and on Maximum Likelihood methods have been added. In keeping with these additions some computer programs have also been added and a few have been modified.