Customer Reviews

Numerical Analysis

5 star:		(1)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

I'm very pleased to see this book still in print. I used it for the few years that it was originally available as a textbook (1988-1990) for a course in numerical methods / computational engineering for beginning M.S. engineering students. It does not provide the omnibus coverage of the usual texts, and it is not a mathematician's text on numerical analysis. It develops ideas with examples, then summarizes with theorems, but these are not written in the formal style, thick with strange symbols, that one finds in a mathematician's numerical analysis textbook. These are good features for engineering students, because it helps the student focus on what is there without being overwhelmed. The presentation is readable: How the methods work, and why you need them is presented with examples and in the context of the sort of problems in which they are needed. I particularly liked the development of upwind vs. central differencing.

The following statement (pg. 165) perhaps best reveals the character of this book:

"Now, convergence theorems are usually _very_ difficult to prove and, often, one finds it embarassingly easy to assume more than one actually needs numerically. This will be the case for the theorems which follow, in order to avoid the need for advanced mathematical theory. We will indicate which these additional assumptions are as we proceed."