Consensus of Open Multi-Agent Systems under Dynamic Event-Triggered Mechanisms

This paper studies the consensus problem in open multi-agent systems with time-varying agent set, where discrete control updates and intermittent communication are employed. An auxiliary dynamic variable is introduced to design an event-triggered mechanism with adaptive thresholds. To address topology changes due to agent arrivals and departures, the network is modeled as a set of switching subsystems with varying scales. Using switched system theory, we establish that exponential consensus convergence with non-Zeno behavior is guaranteed if each subsystem's dwell time exceeds a predefined lower bound and the switching frequency remains finite. Furthermore, a self-triggered prediction method is developed to handle incomplete state information, utilizing historical data from triggering instants to achieve fully discretized communication. Numerical simulations validate the effectiveness of the proposed framework in balancing communication efficiency and consensus performance under dynamic network conditions.