A hybrid Immigrants schema for particle swarm optimization algorithm

The complexity of real-world problems raises new challenges to evolutionary computation. Responding to those challenges, several methods are developed in particular the evolutionary algorithm. They are easy to implement and can provide good results. Among those methods, the Particle Swarm Optimization has some gaps in term of the convergence and diversity. In order to enhance its performance, we propose in this paper a population-distributed model using particle swarm optimization algorithm and island model structure. We investigate the effect of migration parameters such as the topology migration and the migration strategies. The empirical results using set of test functions show that the proposed PSO model is better to achieve an appropriate tradeoff between intensification and diversification.