基于影响力网络模型的多智能体一致性协议

如何增强系统一致性是多智能体系统研究中的一个重要问题。传统一致性协议通常未考虑拓扑中的关键节点，并且拓扑权重单一，从而导致系统更容易分裂。基于人际关系网络中的关键人物可以促进不同社区信息交流的思想，提出了一种影响力网络模型（influence network model，INM）。首先，提出了分布式的Hub Node识别算法（distributed hub node identify algorithm，DHNI），用于区分关键节点和非关键节点，可以应用在分布式多智能体系统中。其次，提出了基于分布式hub node的拓扑权重设计算法（distributed hub node-based topology reweighting algorithm，DHNTR），量化不同节点对其邻居的影响力。最后提出了基于影响力网络的一致性协议。设计了公共Lyapunov函数，分析了系统的全局稳定性，证明了系统具有Lyapunov意义下的稳定性。仿真实验表明该协议可以增强系统一致性。

Influence network model-based multi-agent consensus protocol

How to enhance system consensus is an important problem in the research of multi agent systems. The conventional consensus protocols usually do not consider the hub nodes in the topology, and the topology weight is single, which make the system easier to split. Based on the idea that key figures in the interpersonal network could promote information exchange in different communities, this paper proposed an INM. Firstly, this paper proposed a DHNI algorithm to distinguish between hub nodes and non-hub nodes, which could be applied to distributed multi-agent system. Secondly, this paper proposed a DHNTR algorithm to quantify the influence of different nodes on their neighbors. Finally, this paper proposed an influence network-based multi-agent consensus protocol. This paper designed a common Lyapunov function and analyzed the global stability of the system. this paper proved the stability of the system in the sense of Lyapunov. Simulation results show that the protocol can enhance system consensus.