The Genetic Algorithm Approach: Why, How, and What Next?

When man wanted to fly, he first turned to natural example — the bird — to develop his early notions of how to accomplish this difficult task. Notable failures by Daedulus and numerous bird-like contraptions (ornithopters) at first pointed in the wrong direction, but eventually, persistence and the abstraction of the appropriate knowledge (lift over an airfoil) resulted in successful glider and powered flight. In contrast to this example, isn’t it peculiar that when man has tried to build machines to think, learn, and adapt he has ignored and largely continues to ignore one of nature’s most powerful examples of adaptation, genetics and natural selection. The primary mechanisms for adaptation in most optimization and learning systems depend upon man’s own artificial creations such as calculus and counting. The rich and efficient performance of nature’s own adaptation algorithm-of-choice is just starting to receive the attention it deserves in artificial system adaptation and learning.