Review of How to Solve It: Modern Heuristics

Abstract

How to Solve It is a friendly gem of a book which introduces the basic principles of traditional optimization, evolutionary optimization, neural nets, and fuzzy methods. In the spirit of Polya's classic of the same name, the authors emphasize the "how and why" of the problem-solving process, constantly prodding the reader to stop and solve subproblems, or come up with new heuristics, or indeed question whether or not the problem has been posed correctly in the first place.This book is clear, concise, and fun to read. It is not a handbook of heuristics, but rather an accessible high-level overview of the pros, cons, and applicability of the major categories of heuristic methods, with particular emphasis on optimization methods utilizing evolutionary computation. The presentation is lucid, and the authors do a good job of picking out key properties of algorithms and problem domains. The only prerequisites are basic mathematics and some problem-solving talent.