Migration Effect of Hierarchical Multi-population Genetic Algorithm

This Metaheuristics are easy to use and usually get good solutions for complex problems. There were many approaches proposed in improving the performance of the metaheuristics. In the past, we proposed a hierarchical execution architecture of GA to shorten the generation number needed to find final solutions. In this paper, we embed different migration mechanisms in the hierarchical sub-population architecture and compare their performance. Experiments are also made for verifying the effectiveness and efficiency of the proposed migration mechanisms. From the experimental results, it is found that the migration mechanism in the hierarchical sub-population architecture could improve solution quality.