Customer Reviews

AI Techniques for Game Programming (Premier Press Game Development)

5 star:		(14)
4 star:		(4)
3 star:		(1)
2 star:		(2)
1 star:		(1)

 

 

Beauty is in the eye of the beholder, so let me tell you what I was looking for: a non-academic AI book with practical examples for implementing genetic algorithms and neural nets. I don't have a math degree (though I understand basic algebra and remember a bit of trig), and I learn best by doing hand's on project, not by theorizing.

This book fits the bill perfectly. It is well written, humorous, clear and patient. The examples are interesting enough that you can see how they would be useful for solving other problems, not necessarily game related.

One caveat: if you don't have a decent intro to basic Win32 API programming, get Charles Petzold's Programming Win32 book and get busy. Yes, the first two chapters of AI Techniques are a Win32 refresher (which was good for me because I last wrote Win32 3 years ago...I now do Java only), but I'm pretty sure if you haven't seen it before you'll be lost.

BTW, just to give you an idea how clear the concepts were presented, I've recoded Chapters 3 and 5 as properly OOPed (MVC, etc) Java applications. Just MHO, but Java is a much better platform for this stuff. ...

This book IS my Masters thesis. I built a multi layered combative system driven wholly by neural networks evolved using genetic algorithms. The learning of agents was unsupervised and they existed collaboratively and adversarily. If all of this made no sense, dont worry! Read this book and it will.
The book covers these techniques (except for multi layered architectures) to a level understood by anyone with a basic knowledge of C++. It totally demystifies NNs and GAs. Other books on these subjects actually put you off the entire concepts of AI by feeding you fear and confusion.
I have completed the book and have read it numerous times. It is going to be invaluable for my development of PHd stuff - his writing introduces areas for potential research.
Im would definately buy other books from this author. I hope he covers other areas of AI soon and puts 'em into print.
Be jaysus, tis mighty I tell ya!!!

After reading (and enjoying) the author's web tutorials some time ago I was looking forward to receiving this book - and it hasn't disappointed.

The author has managed to squeeze in a pile of information about GAs and neural nets and yet managed to keep the math down to a minimum - which for a moron like me is especially good news! The source code is kept simple and is very easy to follow.

The example programs that accompany each chapter illustrate each technique very well, and more to the point have given me many ideas to try out with my own projects. The mouse gesture recognition example in particular is a great way of teaching backpropagation, something I had completely failed to understand until I read this book.

Although other reviewers have already pointed out everything positive about this book, I felt I also had to praise it because it's so good. It's a great introduction to neural networks and genetic algorithms. By the way the author should be congratulated for his "never include any code that you don't throughly explain" philosophy. Although it's over 400 pages because of this philosophy, you can skim/skip around 150 of them if you know some win32 programming and basic high school math/physics. The book is so concise and easy to understand that I went through it in a night and came up having a firm idea about the subject. There are also a lot of good ideas to try out throughout the book(i.e. tinkering with the source included etc.)

To put it shortly, this is a great book, it's really worth the money you spend on it.

I spent the majority of the time I was reading this book thinking "This is perfect!" Just about every other book of this general type that I have read assumes I know things I don't, is a lot thicker than it needs to be, is much too simple, or is really boring. I don't think the author of this book wasted any pages at all. Everything was put forth in a concise, easy to read tone, and whenever I came across something I hadn't seen before, it was explained in short order. At the same time, he does assume that you know what you're doing, C++ programming-wise, so I didn't have to skip past anything I already knew. For what it's worth, this book created a lot of enthusiasm and confidence in me.

The examples in the book are great. Each chapter takes you through a different small project that helps you to understand what is going on, as well as how it can be applied to games. And, as a bonus, the projects are actually interesting! I really got a kick out of showing my versions of them to people. I was able to code each of them myself in a day or two as I went, but the author's full source code and executables are included if you just want to take a look as you read.

If you are already pretty familiar with windows programming, you probably won't need to read the first two chapters, but they answered a lot of questions for me, and really I felt like the author knew what I was thinking. Later on in the book, there is a review of transforms and matrix math that I found to be a really good reference and refresher.

As a side note, the author wanted to title this book "Genetic Algorithms and Neural Networks for Game Programming," but his publisher didn't think it was 'snappy' enough. He's very active in the forums on his website, and he and others there have been a great help to me.

This book is a well written introduction to Genetic algorithms and Neural Networks, without the usual dependence on high level math. In just a couple of days, I have gotten a clear picture of how both genetic algoritms and neural networks can be used and implemented. The provided demos are simple enough to understand quickly, and they demonstrate the material well (imagine that--a relevant demo). On top of that, they all worked without modification on my machine.

The book does not cover the more traditional "graph search" types of AI like heuristic search and expert systems (which I think is refreshing). Instead, it focuses on the types of AI that learn or evolve. This is probably a much better approach for modern game development (unless you are writing puzzle games), and I highly recommend this book.

This book is a little jewel hidden in the usually trashy prima press
game development series.. it gives a great, clear, non-formal introduction to genetic algorithms and neural networks in the context of game programming.. and gives you example code in good C++ (brainy aliens storming towards you, a genetic algorithm controlled lunar lander.. tanks that move around avoiding obstacles....) that actually works and is simple enough to be understandable for a newbie. The writing style is lively but not annoying. I recommend this title to anyone who has an interest for the fun side of programming....

As mentioned above, this is THE book for AI. It uses plain language and explains different topics using real life examples. If you need/want to learn AI programming, this should be the first book of your collection.

First of all, I must say that the title of this book is a bit misleading, and if it's not what you're looking for, it may be a mistake to purchase. The book is not about AI in general, but instead about two AI learning techniques that are becoming more and more used in the game programming world. These two techniques are 'Genetic Algorithms' and 'Neural Networks'. The book is about how to program these, and only these, techniques. If you are looking for a more general book on AI with a broader range AI subjects, such as pathfinding, then I would suggest a book from the AI Game Programming Wisdom series.

Having said that, this book is wonderful. Not only does it easily explain how genetic algorithms and neural networks work, it shows you step by step how to implement the techniques into your own games. It can also serve as a rather nice introduction into these subjects, as you may encounter more technical explainations of them in other texts. The CD that is included also contains many useful demos and ready-to-use C++ source code, making the examples mentioned in the book quickly accessible.

If you're thinking about buying this book, make sure that it's genetic algorithms and neural networks you want to learn more about. If it is, this book is rather thin but extremely resourceful.

The only other thing I didn't like about this book, besides the title, was the author's decision to spend the first 85 pages of the book talking about Windows Programming. I believe such a thing should not be taught in such a book, because other books can be consulted to learn more about Windows Programming. I don't think this should take a star away from the book, however, since the rest of the book is extremely well-written and resourceful.

I really don't see why so many people freak out and pretend to be all upset because it covers use of the GDI at the beginning. I'll agree that in many cases it's not particularly helpful, but I don't see why people pretend to be incapable of going past those chapters. Anyway, I like this book a lot because it gets you making things that actually work, and it doesn't do it in a way that it's impossible to figure out what the heck the author is talking about. I only rate it 4 stars because it would be better titled as "Neural Nets and Genetic Algorithms for Stuff." It doesn't cover anything other than neural nets and genetics, and it certainly doesn't seem to be limited to games.