Genetic Algorithm Adopting Selective Virus Infection

K. Tamura, A. Mutoh, S. Kato, and H. Itoh (Japan)


genetic algorithm, virus theory of evolution, combinatorial optimization, selective infection


The genetic algorithm have been demonstrated its effectiveness in various optimization problems these days. But GA has problems that lose populations diversity in early generation called early convergence and that genetic operations called crossover and mutation needs a lot of times to spread schema among the populations. GA is based on Darwinism. However, some types of evolutionary hypotheses have been proposed. Virus theory of evolution is one of them, and the Virus Evolutionary Genetic Algorithm based on that theory is reported. In this system, exploitation is executed by coevolution of two populations. On the other hand, difference of searching ability of crossover methods on the first and latter half generation is reported. This denotes selecting crossover method is important to search optimal solution efficiently and steadily. It’s believed that virus infection has influence to searching ability on the each generation too. So we propose one method to select infection’s method, and confirm the effective of our system by experiments.

Important Links:

Go Back