Wyniki wyszukiwania - BazTech

Ograniczanie wyników

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Efficient selection methodsin evolutionary algorithms

Stańczak Jarosław

Computer Science

2024

T. 25 (1)

95--122

Evolutionary algorithms mimic some elements of the theory of evolution. Thesurvival of individuals and the ability to produce offspring play significant rolesin the process of natural evolution. This process is called natural selection.This mechanism is responsible for eliminating weaker members of the popula-tion and provides the opportunity for the development of stronger individuals.The evolutionary algorithm, an instance of evolution in the computer envi-ronment, also requires a selection method – a computerized version of naturalselection. Widely used standard selection methods applied in evolutionary al-gorithms are usually derived from nature and prefer competition, randomness,and some kind of “fight” among individuals. But the computer environmentis quite different from nature. Computer populations of individuals are typi-cally small, making them susceptible to premature convergence towards localextremes. To mitigate this drawback, computer selection methods must in-corporate features distinct from those of natural selection. In the computerselection methods randomness, fight, and competition should be controlled orinfluenced to operate to the desired extent. This work proposes several newmethods of individual selection, including various forms of mixed selection, in-terval selection, and taboo selection. The advantages of incorporating theminto the evolutionary algorithm are also demonstrated, using examples basedon searching for the maximumα-clique problem and traditional Traveling Sales-man Problem (TSP) in comparison with traditionally considered highly efficienttournament selection, deemed ineffective proportional (roulette) selection, andother classical methods.