Distributed Optimization-Based Formation Control: A Dynamic Event-Triggered Approach

This paper investigates an event-triggered approach for distributed optimization-based formation problems of multi-agent systems (MASs) under directed graphs. A distributed optimization-based algorithm is proposed such that agents can rendezvous around an expected time-varying formation, and at the same time, optimize a global objective function. A novel dynamic event-triggering mechanism is proposed, where a buffer variable is introduced to store the historical local information including the average values of measurement errors over one transmission interval. Then, the closed-loop MAS is proved to be input-to-state exponentially stable with respect to external inputs. Zeno-freeness is guaranteed by a computable minimum interevent time. Moreover, the trade-off between network load and computation complexity is illustrated by comparing two kinds of event-triggered mechanisms based on different signals as inputs to the buffer variables. The effectiveness of the proposed method is verified by numerical examples.