Genetic Programming: On the Programming of Computers by Means of Natural Selection (Complex Adaptive Systems) , by John R. Koza

Oh, the monster of a book! If you want to learn to do genetic programming, then this is the book for you. If you need an interesting presentation of what genetic programming is, then this book is way too heavy.

Let's start with the beginning. Genetic Programming: On the Programming of Computers by Means of Natural Selection (Complex Adaptive Systems) is a scientific book written by John R. Koza to explain why, how and what to do to make your computer find solutions to problems by using natural selection algorithms to automatically create programs to solve them. This is not a new field and a lot of research has been done in it, but this book takes it almost to the level of encyclopaedic knowledge.

First, Koza submits the idea that genetic programming can be used in most problems where computers are been used. That's a bold claim, but he proceeds on demonstrating it. He takes problem classes, provides code to create the programs that solve them, shows results and statistical analysis on the results and explains what the algorithm did to create said program at specific iterations. That's a lot to take in. If you are working on a program and you are using the book, you are more likely to find it extremely useful, both as a source for information and as a reference that can always be consulted.

However, if you are a casual reader like myself, reading all that code and statistical analysis in the subway can be difficult. And it's a lot of book, too. So, after some consideration, realising that I have no current project on which to apply the knowledge within the book, I've decided to stop reading it. I got to about a quarter of it, so I can safely say that it is a very thorough and well written book. You just have to need it in a certain way.