Theoretical overview: why and how does it work?
Genetic and evolutionary algorithms

The big question is when and how to use AEs effectively, since they are optimization algorithms that are quite complex to implement and relatively costly in terms of computing time (as mentioned above, the "black box" approach is not recommended). The answer to this question is not simple. However, numerous theoretical and experimental analyses provide some insights into when the use of an AE is justified or can offer advantages in comparison with or in combination with conventional techniques.

Historically, pattern theory represents the first global convergence theory for genetic algorithms; it is in fact an extremely simplified model of the behavior of an AE