Consensus of hierarchical multi-agent systems with a time-varying set of active agents

Abstract: Time-varying hierarchical multi-agent systems are common in many applications. A well-known solution to control these systems is to use state feedback controllers that depend on the adjacency matrix to reach consensus. This solution has been applied so far to multi-agent systems with fixed or time-varying communication topologies. In this paper, we consider single-integrator multi-agent systems where only a subset of the agents are active at any given time and the set of active agents is time-varying. This type of multi-agent system is relevant in applications such as modular multilevel converters and water pumping systems. We develop a switching algorithm that periodically selects a set of active agents through a set of graphs that connect the follower agents with a leader agent. We further prove that the developed switching algorithm combined with a classic consensus protocol yields convergence to a consensus state defined by the leader despite a time-varying set of active agents. The effectiveness of the switching algorithm is illustrated in two benchmark examples: a modular multilevel converter and a water pumping system.