Genetic algorithm

A Genetic algorithm (GA) is a computer-based metaphor of Darwin's theory of natural selection.

In such algorithms, an individual represents a set of parameter values.

Description of the Simcenter Amesim GA

• The first step of the GA is to randomly generate a population of individuals.

• Then the best individuals are kept whereas the others are removed from the population and replaced by "children" of the others. Children are obtained by randomly choosing two "parents" among the best individuals and choosing parameters close to the parent ones.

• In addition to this, individuals mutate: their characteristics (parameter values) are changed by adding perturbations to their values.

• After several generations, individuals converge to one or several best solutions.