Making Genetic Algorithms Fly

The interlocked goals of this book—the design of GAs that work and the discovery of computational models of innovation—force us to ask a pair of methodological questions: How do we design competent, selectorecombinative genetic algorithms—GAs that solve hard problems quickly, reliably, and accurately? How do we build computational models of the processes of cross-fertilizing innovation? If we were to treat these complex questions about abstract topics directly with the intellectual gravity and abstruse prose that they perhaps deserve, there would almost certainly be no readers left at the beginning of chapter 3. But more to the point, it is often useful when trying to do something new to shift the terms of the debate to a domain that is further along in its intellectual development. Thus, here we choose not to examine the question of how we should design GAs directly; rather, we shift to another era and area of human endeavor and examine the historical record of how two of the greatest inventors of the twentieth century came to build and fly that marvelous machine, the airplane. Along the way, we abstract a number of straightforward lessons that will help us in the development of innovating machines such as genetic algorithms.