Distributed Optimization Design of Multi-Agent Systems With Packet Losses
Shenghuan Chen, Weiyao Lan, Ji Ma, Xiao Ping Yu · 2020
In this paper, we investigate a distributed optimization problem of the continuous-time multi-agent systems (MASs) with packet losses. We first prove the boundedness of the state in the presence of packet loss, which means that the gradients of objective functions are not required to be bounded. On this basis, we propose a sampled-data distributed dynamic scheme such that the all agents' states can converge in probability to an exact optimal point. Furthermore, if we know the probability of random packet losses in advance, the upper bound of the sampling interval is derived. An illustrative example is given which substantiate the usefulness of the proposed method.