Clustering as a means of leader selection in consensus networks

Abstract: In the leader-follower approach, one or more agents are selected as leaders who do not change their states or have autonomous dynamics and can influence other agents, while the other agents, called followers, perform a simple protocol based on the states of their neighbors. This approach provides a natural link between control theory and networked agents with their input data. Despite the fact that the leader-follower approach is widely used, the fundamental question still remains: how to choose leaders from a set of agent. This question is called the problem of choosing leaders. There is still no selection algorithm that is both optimal under a natural criterion and fast. In this paper, for agents that obey a linear consensus protocol, we propose to choose leaders using graph nodes' clustering algorithms and show that this method is the most accurate among the fast existing algorithms of choosing leaders.