Papers by Keyword: Population Diversity

Abstract: A new particle swarm optimization algorithm (a diversity guided particles swarm Optimization), which is guided by population diversity, is proposed. In order to overcome the premature convergence of the algorithm, a metric to measure the swarm diversity is designed, the update of velocity and position of particles is controlled by this criteria, and the four sub-processes are introduced in the process of updating in order to increase the swarm diversity, which enhance to the ability of particle swarm optimization algorithm (PSO) to break away from the local optimum. The experimental results exhibit that the new algorithm not only has great advantage of global search capability, but also can avoid the premature convergence problem effectively.

Abstract: HFC (Hierarchical Fair Competition) model is a kind of sustainable EA based on hierarchical search. On the basis of studying the sustainable evolutionary theory of HFC model deeply, in the HFC model the admission and export threshold are proposed to divide all fitness grade. And it is suggested that both static and active thresholds is used by the complication of different problems. In addition, through analyzing the parameters and topology space of HFC model, it is gained to the organize framework and detailed algorithm structure. Finally, it is proved by exercise that HFC model can maintain the diversity of population and assure the fair competition during individuals effectively.