Selection/ selection operator

necessary operator in EA, which, depending on the fitness or the objective function values, respectively, decides on the reproduction of the individuals. Selection gives a direction to evolution, it guides the search (e.g., in the sense of optimization or amelioration). It conserves successful states, but at the same time it reduces the diversity of the population. In this sense, selection is the antagonist to variation, which produces diversity. If the selection method guarantees the conservation of the information of the best individual, it is called ``elitist''. Popular selection methods are truncation selection with the variants ``$+$''- and ``$,$''-selection, tournament selection, fitness-proportionate selection (realized as roulette selection or stochastic universal sampling) and their transformed versions with ranking (ranking-based selection), as well as Boltzmann-selection. All of these selection methods reinsert the individuals into the current population after selection, i.e., individuals can be selected more than once.