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

It has been studied that a few informed individuals in a group of interacting dynamic agents can influence the majority to follow the position and velocity of a virtual leader. Previously it has been shown that a cluster-based selection of informed agents can drive more agents to follow the virtual leader compared to a random selection. However, a practical question is: How many informed agents to select? In order to address this, here we propose a novel method for selecting informed agents based on community structures in the initial spatial distribution of agents. The number of informed agents are decided based on the strongest community structure. We test and analyze the performance of the proposed method against random and cluster-based selections of informed agents using extensive computer simulations. Results of our study show that community-based selection can be useful in deciding an optimum number of informed agents such that a majority of the group can achieve their common objective.