Realization of island model genetic algorithm on Nvidia Cuda architectures

Realization of island model genetic algorithm on NVidia CUDA architectures is described. Analysis of numerical efficiency of graphic processor units is made. Features of CUDE using exactly for genetic algorithms and for island model practically is shown. An island model suitable for CUDA is designed. Parameters of genetic operators for such model are précised. Scopes of CUDA applying for such tasks is determined