At the forefront of Artificial Intelligence

AI Techniques for Game Programming

Cover 9.5

Author: Mat Buckland

Excerpt:
AI Techniques for Game Programming gives you the keys to incorporating artificial intelligence into your games. Three sections walk you through basic Windows® programming, the theory and practice of genetic algorithms, and finally neural networks where you can learn about the biological makeup of real brains and behavior evolution.

.: Buy at Amazon.com :. .: Buy at Amazon.co.uk :.

Targeted Audience for this Review:

This review is targeted for beginners. I myself am a novice in the field of AI and I do not have years of experience or diversity in my readings to provide a more critical review. However, it is in this benefit that others who are new to the scene can read what I have experienced without a bias evaluation.

Introduction

The first thing that naturally comes to mind when starting out in the world of AI is "Where do I start?" It is a very important question and if not answered properly you could be left in a sea of confusion. More importantly, you could lose total interest in AI. This book helps bridge the gap between the complexities of genetic algorithms and neural nets to something anyone can understand.

Content - 5/5

Part 1 - Win32 API

This section provides you with an understanding of the Windows environment and how you can develop programs with the Win32 API. It guides you on how to create windows, make efficient use of resources, display graphics, and setup double buffering for smoother animations. This book has done an excellent job in giving you all the details in just under two chapters. Most books would go into hundreds of pages of boring detail, but this one gets you straight into the action.

Part 2 - Genetic Algorithms

This section of this book discusses genetic algorithms. It starts with a description of basic biology, to how it is represented in software, to using it in evolving your programs over time. It provides a step-by-step basis to getting you familiar with genetic algorithms, their use, and even gives you extra techniques to further improve the results of your programs or experiments.

Part 3 - Neural Nets

This section provides information on how to develop neural nets. It gives you a good explanation of what they are, how they are used, and techniques for further improving the results you obtain. It explains to you how genetic algorithms can develop evolutionary neural nets and how you can train your neural nets through a supervisory technique to recognize patterns.

Source Code and Executables

The book not only teaches you the topics at hand, it also provides you with source code and executables for every chapter to test out what you just learnt. The source code is well commented and easy to understand. The only pitfall is that you need to understand C++ if you are to read the source code.

Overall Content Score - 5/5

The content is definitely there. With the topics discussed, combined with source code and executables, you can enjoy best of both worlds.

Writing Style - 4/5

The writing style is very relaxed. The author tries to explain things to you as you where there in person. This I found was a very effective approach because it makes you comfortable with what you read and boosts your confidence if you have any doubts. Coupled with the book are short stories and real world facts that help further your understanding and interest in the topics at hand.

The only pitfall here is that the chapters you read supply source code examples in C++. This is to help eliminate any confusion and show you exactly how it looks in software. Though, without knowledge of C++, it would make your effort to learn genetic algorithms and neural nets a little more difficult.

Each chapter also includes a set of exercises for you to complete. These can be helpful for anyone wanting to test their knowledge of what they just read. It also encourages you to play with the source code so you understand the requirement of experimentation with AI.

As an additional note, if you read the articles on the more theoretical side of genetic algorithms or neural nets here on Generation5, you might be left with some confusion. When you are done reading this book, you will be able to look back at those articles and laugh at how easy they are to understand.

Overall Writing Style Score: 4/5

I didn't give it a perfect score because of the language the source code was provided in. I don't have any quarrels with C++. In fact it is my preferred choice of language. Those without C++ knowledge (say Visual Basic or Delphi) are missing out big time and would be leaving much to desire by skipping the example source code.

Price - 5/5

Seriously, it actually scares me to write this thinking that by chance the price of the book might increase. It is an amazing price for what you get in return. Amazon provides you with a nice discounted price (compared to Chapters) and was well worth the money. To this date I have found no better substitute that could teach me genetic algorithms or neural nets as easy as this book has.

Additional Support - 5/5

If reading the book isn't enough, the author also has his own website with primers on genetic algorithms and neural nets. The online tutorials are just a savory taste of what the book actually covers. The link is included at the top of this review.

On top of having his own website, the author is also active in the forum community by providing support to people with questions and participates in AI conversations. It is a unique opportunity to actually converse with the author of the book you read.

Conclusion

Give a man a fish and he will eat for a day. Give a man the sea and he will eat for an eternity. This book doesn't just feed you what you need to hear. It provides you with insight as to how you can further your knowledge though experimentation. It is through experimentation that gets you to perfect your AI, especially if you are to develop AI for games.

I was very impressed with the book. If others would like to know of my success with the book, please check the AI section of my website for a list of completed demos.

Overall Score - 19/20 - 95%

By now you have read my pros and cons of the book. If you know C++, don't hesitate on such a good deal to get you started into AI.

Review Summary:
Content	5/5
Writing Style	4/5
Price	5/5
Additional Support	5/5

Pros:
- Easy to read and understand - Affordable price - Great introduction into AI
Cons:
- Source code provided only in C++

Submitted: 26/07/2003

Article Toolbar

BibTeX entry