"Genetic algorithms are a category of optimization techniques, but looking at fitness we can see our problem: first it is attempting to solve a search problem (finding for working solutions, of whatever quality), and then trying to optimize (improving the quality of discovered solutions). Very few--if any--algorithms are going to perform well for both the search and optimization aspects of this problem "

Interesting article with lots of examples in python.

http://lethain.com/entry/2009/jan/19/genetic-programming-a-novel-failure/