Evolutionary algorithms (EAs) are a biologically inspired set of rules that describe the use of an evolutionary process in computation. In the typical EA, during a generation, members of a population are ranked according to a fitness function. Those members with the highest fitness ranking are given a higher chance to become parents for the next generation, the offspring. The exact method used to generate offspring from the parents, is termed the reproduction heuristic. The common reproduction heuristic is a mutation rate (per chromosome) and sometimes also a chance of having mutations. After a number of iterations, a typical population increases its fitness and converges towards (local) optima. EAs function very well at optimization when the search space is large.
Erik de Bruijn