Cluster-based informed agents selection for flocking with a virtual leader

Abstract

Recent literature on flocking in multi-agent systems show that a minority of informed agents in a group of dynamic agents can influence a majority to follow a virtual leader. However, it is not reported how to select these informed agents from the group in order to increase the number of agents which will eventually follow the virtual leader. In this paper, we propose a cluster-based informed agents selection method to achieve this objective. The proposed method enables us to select informed agents such that they are spatially evenly distributed within the group of agents. We carried out extensive simulations to analyze performances of the proposed method against the traditional random-based informed agents selection method. Simulation results show that the proposed method can increase the number of agents which eventually follow the virtual leader for a given number of informed agents. Therefore, the proposed cluster-based informed agents selection method is useful for leading the majority of a group with less number of informed agents.