Review of "Linear Programming: Foundations and Extensions", Robert J
Vanderbei
Kluwer Academic Publishers ISBN: 0792398041 92.50
Vanderbei has produced a splendid book. Splendid, but not quite perfect
yet, and with small revisions and a rational pricing strategy from the
publisher it will be possible to recommend the second edition without
reservation.
He deals at a fairly elementary and introductory level with the simplex
method, network problems, interior point methods, integer programming,
quadratic programming and general convex programming. Along the way
there are some detours into convex analysis, elementary game theory,
regression and the optimisation of structures.
Implementation issues are covered at a level that would interest the
inquisitive reader yet not burden him or her with unnecessary details.
For instance sparsity - vital to any practical implementation - is
covered very nicely indeed and numerical stability is emphasised.
The section on interior point methods is as beautiful and elegant as one
might have hoped from one of the major contributors to the area. It is
the best presentation I have seen and it would be very easy to teach an
undergraduate course from this book with the expectation that
the students would get a lot from the material. Providing of a lot of
relevant software on the book's Web site is an added bonus.
The author's wit shines through in the text (quite why he dislikes
Newark so much is not clear!) and the selection of examples has been
done with great care.
The worked examples are excellent and very much to be welcomed.
Resorting to clear examples to demonstrate the theory is usually
deprecated - here it is done to great effect.
I have a few quibbles. There are occasional lapses into some
mathematical terminology where it is unnecessary. There is one piece of
unexplained notation which I and everyone I have shown the book to have
never seen before. The coverage of post-optimal analysis is too brief
and I think the chapter on optimising structures is slightly self
indulgent (but why not?). There is nothing on decomposition.
My major complaint is the price. Kluwer must have no idea what an
undergraduate is prepared to spend on a book. Let's have the second
edition in paperback at a sensible price please. Then this book will
become a classic.