AI Techniques for Game Programming

by Mat Buckland

"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 examples to be found in other books. Each chapter takes readers through the theory a step at a time, explaining clearly how they can incorporate each technique into their own games. After a whirlwind tour of Windows programming, readers will learn how to use genetic algorithms for optimization, path-finding, and evolving control sequences for their game agents. Coverage of neural network basics quickly advances to evolving neural motion controllers for their game agents and applying neural networks to obstacle avoidance and map exploration. Backpropagation and pattern recognition is also explained. By the end of the book, readers will know how to train a network to recognize mouse gestures and how to use state-of-the-art techniques for creating neural networks with dynamic topologies.

Genres: Programming, Software, Computers, Artificial Intelligence, Nonfiction

448 pages, Paperback

First published October 14, 2002

Reviews

[Tom · Rating 4 out of 5]

I bought this book when it came out and wasn't quite ready for it. Here we are many years later and I'm digging into it again to see if my older mind can comprehend this any better. The concepts do seem to be clicking now and I think converting the code from Windows C++ to GDScript (to work with Godot) actually seems to help me understand the information even better. I can't just blow past a concept because I need to reinterpret the code. Don't let the Windows centric slant prevent you from picking this up if you have an interest in the subject.

I also enjoyed the friendly writing style as it's not as dry as other books on the subject can be.

[Wayne · Rating 4 out of 5]

Mat has a great way of explaining what could be a complicated subject and provides very good examples. He provides good samples that really get you up to speed with the basics (and common) areas of the technique to the point that at the end of the book you should be able to approach a different scenario in an effective way.

The only criticism I could make is that you need to have some understanding of Windows GDI programming. This is an area where the book is showing it's age a little and come people may not be too familiar with it. Having said that though, there is enough wisdom packed into this book that it's worth it just for the theory.

Overall a good book for understanding the concepts of neural networks and genetic algorithms in the context of gaming.