Simulation platform for flocking in multi-agent systems with a virtual leader

In recent years, the broad applications of multi-agent systems in many areas have stimulated a great deal of interests in studying flocking problems. In this paper, we design an animated simulation platform for flocking by using Java language, which can show the users practical effect of flocking algorithm with a virtual leader. The platform includes three modes. In the first mode, we assume that every agent in the group is an informed agent, and use Olfati-Saber flocking algorithm for this mode. In this mode, all agents in the group can track the virtual leader. However, only fraction of agents are informed agents in most nature examples and engineering applications. In the second and third modes, we both assume that only fraction of agents are informed agents, and use the flocking algorithm with a minority of informed agents. There are two different strategies of selecting informed agents in the group for the second and third mode. Informed agents are randomly selected by the second mode, and specifically selected by the third mode. Simulations for both modes show a very small fraction of informed agents can make most of agents track the virtual leader.