| Resumen |
Cellular Genetic Algorithms (cGAs) are a special kind of Evolutionary Algorithm (EA) that perform heuristic search on the style of Genetic Algorithms but over highly structured populations. Their distinctive feature is to organize the search space with a neighborhood structure that limits and governs the evolutive relations among individuals. Each evolving individual can only interact with its designated neighbors, thus forming a connected graph that twists the development of the reproductive and mutation processes. The intrinsic parallel structure of cGAs can be exploited and forward sttudied using Graphical Processor Units (GPU) and this paper shows the implementation techniques as well as the emerging conclusions of that performance study. |
