Customer Reviews

How to Solve It: Modern Heuristics

5 star:		(15)
4 star:		(2)
3 star:		(2)
2 star:		(1)
1 star:		(0)

 

 

This book provides a very accessible and contemporary treatment of optimization. Of particular interest is the problem solving orientation of the book as opposed to a tool-based approach to optimization and heuristics. The writing style of the book makes the book very interesting and readable - a rare thing to say about technical books! I used this book in a Master's class on Heuristics (Systems Engineering, University of Virginia) and received the most positive textbook reviews I have seen in my fifteen years of teaching. The book is an excellent choice for a course on heuristics, mathematical modeling, optimization, etc., and could be used in an advanced undergraduate class or a graduate class. In addition, the book is ideal for practitioners who may not have had exposure to modern heuristics in their education or practice, or those who want to get updated on the latest developments in the field.

I read this book while taking an advanced class in heuristics. I found the book to be extremely well written and very compelling to read. Although dealing with advanced topics, the authors' friendly and clear writing style makes it accessible to anyone with a CS background.

The first half of the book is on search heuristics, covering methods such as traditional searches (exhaustive search, greedy algorithms, divide and conquer, dynamic programming, A*, etc), methods to escape local optima (simulated annealing, tabu search), and, perhaps most interesting of all, evolutionary algorithms. I later found out that these topics are typically taught in undergraduate artificial intelligence courses, an elective I never took. The second half of the book covers even more advanced areas, such as contraint-handling, neural networks, and fuzzy systems.

The authors use three recurring example applications to demonstrate each search technique: the boolean satisfiability problem (SAT), travelling salesman (TSP), and a nonlinear programming problem (NLP). I really liked the consistent use of these three examples, as they give a sense of continuity throughout the book that helps the reader compare search techniques clearly. I had of course studied the TSP problem in my undergraduate algorithms class but never in the context of such interesting approximation algorithms. In my heuristics class we had assignments to implement the TSP search problem using the Lin-Kernighan method, dynamic programming, and an evolutionary algorithm.

The written English in this book is simply outstanding and crystal-clear, which was something of a shock since I was unable to even pronounce the first author's name. The writing is in a very friendly tone with elements of humour dispersed throughout. Interestingly, in the summary chapter, there is an anecdote on the 1980s TV show Magnum PI (I even remember the mentioned scene myself), further revealing the friendly, plain-English tone of the book. Perhaps the best part of the book is that the numerical mathematical discourse is kept at a minimum (used largely for the NLP problems), so people who haven't taken calculus in ages (like me) can easily enjoy the book.

As an added bonus(!), between each chapter is a brain-teaser problem like those found in those legendary Microsoft interview questions.

My only complaint is that there is no simple analysis of the running time complexity of each algorithm, which even in its simplest form would have been a great thing to read about.

In summary, this book is an excellent read if you enjoy the topics covered. Highly recommended.

This book provides one of the most comprehensive views of modern techniques in problem solving. The authors use a number of classic problems to illustrate conventional heuristics as well as giving you a solid and working knowledge of more modern evolutionary techniques. The appendicies provide a good introduction to background information on probability theory and statistics used throughout the book, as well as projects for further exploration. Scattered throughout the text are complete and up-to-date references that can be used by the reader to delve deeper into certain topic areas. This book is written to be read and understood by both students and experienced researchers in the field.

READ: this is not just another optimization book! Instead of spoon-feeding one technique after another (do a search on "optimization" and you will know what i mean), it challenges you to think CREATIVELY. It says, "if you have a hammer, everything looks like a nail." Read and find out why the more textbooks you read, the more a screw looks like a nail! (and remedy to return to reality)

Despite working on algorithms for years in graduate school, for the first time there is a book that looks at problem solving with a fresh, unbiased perspective. Definitely my best buy in years.

Don't think that this book is just another version of numerical recipes or "how to" for optimization methods. For me it is about something absolutely different. About breaking old, bad habits in problem solving and looking for the simplest and the most elegant solutions for the given problem. Sometimes it will be something complicated, like competitive neural network, but sometimes the solution will be just: "let's assume that there's no river" (see page 185 of this book). Don't put artificial intelligence where just the common sense will be absolutely enough. I remember some of the problems presented there from my high school years. I had more problems with solving them today than it was many years ago. It looks that we are loosing somewhere, in the process of education, the possibility to simplify problems and rather try to solve them by "brute force". This book may give you this fresh look again (I hope).

I first ordered this book thinking it was George Polya 's book "How to solve it", then I realized it wasn't and I bought it anyway since I thought it might turn out as a "must read" book, just like Polys'a book.

One one hand it was a dissapointment, because the books are not written in the same manner and don't attact similar problelsm.

But then, this book makes you look into problems, and realize that usually we people are usually good in solving problems of the sort we learned how to (well... duh!), but surprisingly, we have a hard time solving even trivial problems if they are not placed in the context we got used to seeing them.

This book comes and tries to make things better in this department, showing you some general methods for solving problems, and also showing problems and suggested solutions along with a long discussion.

You should be able, once you've read the book and put your mind to it, to be better in understanding problems, understanding which tool to use for solving them and finally, understanding the tools enough to be able to actually solve the problem.

I enjoyed the overview of methods, and there are many such methods throughout the book (perhaps a complementary book for learning which "machine learning" methods are available these days and what sorts of problems they are useful for solving would be Tom Mitchell's "Machine Learning" book).

I wasn't sorry for buying this book. I'm happy I was fortunate enough to bump into it.

This is simply the best book on computer problem solving that I've seen. I have both editions. The second is expanded from the first with new material on things like multicriteria decision making. The book's engaging tone is matched by a detailed understanding of all the different approaches to problem solving that are offered. The text emphasizes evolutionary computing but offers complete treatments of other optimization methods, although there are only single chapters on neural nets or fuzzy logic. Most importantly, it does so in a way that no other book I've seen does -- it makes it fun and it makes you think! I saw that another reviewer said the book got great classroom reviews. I don't doubt it. I wish there was a book like this when I was in college.

I am a computer scientist, but have gotten impatient over the years with the needless formalization that occurs in algorithmic texts. This is a delightful breath of fresh air in terms of balancing erudition with attempts to be "user friendly". If you want the latest and greatest twist to a well known technique, this book won't provide it. But it does a great job of competently and lucidly explaining the value proposition behind each optimization method and how to gradually upgrade from applying it naively to the more intricately optimized applications. Well done!

The authors have updated their successful first edition, though the latter, printed in 99, was scarcely obsolete. A heuristic can be basically a rule of thumb, dressed up in fancier language. What the authors intend is for you to develop an intuition about when to use modern algorithms. Where is almost every case, these are actually implemented on a computer; a reflection of the cheap availability of computing power to most readers.

The book is a good complement to various standard algorithm texts, like those by Sedgewick, Aho and Knuth. You can consider this book as standing a level above those. [Though Knuth's books also do an excellent job of suggesting when to use or modify algorithms. ]

The level of discussion here is not of a strict, heavy mathematical approach. It can be read as informal guidelines, that discuss the gist of such ideas as simulated annealing and evolutionary methods. There is a wide range of example problems, to motivate you in understanding what might be used to solve them.

Having read a couple of very positive reviews, I was looking forward to this, but now that it is here, and I've had a chance to read it, I'm very disappointed. It strikes me as pretty shallow, and it definitely takes the name of Georg Polya in vain. (Michalewicz also writes 'business-oriented' books on decision support - you can tell). It certainly has little, conceptually, to do with Polya's 'How to solve it' (in fact, given the complete lack of any formal theoretical development, the authors are lucky that the man is safely dead). A more accurate title would be 'a bunch of stuff on optimisation, mostly about genetic algorithms and traveling salesman problems, but with a bit on neural nets and fuzzy logic thrown in'. These three technologies used to get sexy articles in the popular computer press about 10 to 20 years ago. It is interesting that the three always seem to crop up together, but they do - or at least they did.

Anyway, the core agenda, which is not heuristics, does poke out at various points. On page 190 there is a revealing passage bout the elusive 'Holy Grail' of 'a perfect evolutionary algorithm for the TSP [Travelling Salesman Problem]'. Now, the world in general would be fascinated by a polynomial solution to the TSP, but the world in general - sorry to say - doesn't actually give a toss if that solution is evolutionary.*

As I said, I was unhappy about the complete lack of real theoretical background which would put any of the discussed methods in perspective/context. The discussion of simulated annealing, for instance, is absent any of the underlying (and powerful) intuitions from statistical physics which, if nothing else, makes the technique much richer, and not conceptually comparable to, tabu search, with which it is discussed in parallel. As far as I can see, the latter is an isolated hack - empirically it may be effective in some applications, but it is not part of a larger conceptual framework. At least if it is not an isolated hack, then the authors provide no evidence - I note that it gets all of three lines in Russell and Norvig.

More seriously, there isn't even any well-founded discussion of mathematical models of evolution. All that I could find was essentially a citation - not even a discussion - of the 'no free lunch' theorems. No Maynard-Smith, no Kondrashov, not even Hopfield's '78 paper - though other later stuff by Hopfield is cited. And without this - and without a lot of other formal theory stuff as well, if I were being honest - there is no hope of a methodological framework. After all, computers are, unavoidably, formal machines. In the end, the evolutionary models that are discussed are not a lot more than a bunch of gadgets, and the authors are reduced to saying that for your own problems, you are going to have to think up your own gadgets. This, to be blunt, is why evolutionary computing remains a niche research area.

A separate problem is the lack of any perspective w.r.t. other, today more commonly used, technologies that address similar problems. The discussion of neural networks and pattern classification does not mention, e.g., that the benchmark classifier technology today is vector support machines, which substantially outperform neural networks. Or that the standard technology for exploring complex function spaces is Monte-Carlo analysis (strictly, a lot of the techniques that are discussed in the book _are_ actually Monte-carlo methods of one sort or another). This does not mean that the methods discussed are _not_ interesting, just that without a perspective, it is difficult to say whether they are appropriate tools for a job.

In the end it was not clear to me who this book is really aimed at. It is certainly not aimed at me. Senior undergraduates in something like operations research might be a target, but I personally would not use it for a CS or applied math class. And I cannot, honestly, forsee it being a lot of use to me outside the classroom.

*Which anyway, a priori, seems unlikely, since idealised recombinatory evolutionary strategies show rapid information gain (N^1/2) in the size of the genome in suboptimal situations, but near the optimum, parthenogenesis with mutation is a better strategy.