Evolutionary Computation: Toward a New Philosophy of Machine Intelligence

Ignore the gushing blurb on the back cover about the book having the latest tools and techniques to let computers learn. While the methods are indeed state of the art when they were written, true learning by computers is still elusive. But so long as you keep that reality in mind, the text can indeed be useful. We see the span of ideas in evolutionary computing. Aided in no small part by the massive and continued increase in computational power. Fogel laments that the book's ideas are still typically outside what is generally taken to be Artificial Intelligence. The book strives to be both a text for researchers and for students. Though of course the two groups overlap. For researchers, each chapter has a long list of references to journal papers and monographs, so that you can go directly to many of the original sources. While for students, the chapters come with a non-trivial set of exercises, that usually involve some programming.

This is an introductory text useful for teaching at graduate levels in Computer/ Information Sciences. The first two chapters provide an overview of the subject and its relationship with other relevant areas. Chapter 4 covers the analysis of the GA with special emphasis to convergence of the algorithm.This is the main chapter of the book. The presentation style of the book is very beautiful.The book should be read by everyone interested in the disciplines of genetic algorithms and/or soft computing.