AI Techniques for Game Programming (The Premier Press Game Development Series)

By Mat Buckland

Book Overview

'AI Techniques for Game Programming' takes the difficult topics of genetic algorithms and neural networks and explains them in plain English. Gone are the tortuous mathematic equations and abstract... This description may be from another edition of this product.

• Format: Paperback
• Language: English
• ISBN: 193184108X
• ISBN13: 9781931841085
• Release Date: October 2002
• Publisher: Cengage Learning PTR
• Length: 448 Pages
• Weight: 1.78 lbs.
• Dimensions: 9.0" x 1.5" x 7.3"

Customer Reviews

Buckland DE-jargonises the field of AI

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!!!

Perfect for me

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.

Not just for games

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. ...

Well written and focused

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.

Great Stuff

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.