Consensus seeking in large-scale multi-agent systems with homogeneous connections by incorporating two-hop neighbor states

Abstract

The development of multi-agent consensus raises the importance of network topology. As the number of agents increases, multi-agent systems (MAS) in a large-scale and high-density topology demand higher resources, which consequently degrades efficiency of consensus. Existing approaches that consider only direct point-to-point neighbors may overlook potential topological information, further hindering consensus performance. To achieve fast consensus in large-scale and high-density topologies, a framework named Homogeneous Connections Based on Agents State Fusions MAS (HCASFMAS) is proposed. The framework extracts broader topology information of consensus degree by fusing states of two-hop neighbors. Leveraging homogeneous idea, agents establish homogeneous connections with neighbors that exhibit a higher consensus degree, ultimately accelerating the consensus process while preserving connectivity. First, a neighbor selection strategy based on consensus degree of agent state fusion is introduced to construct candidate neighbors, aiming to reduce redundant connections. Second, an adaptive consensus algorithm is formulated to flexibly adapt to the distribution of neighbors. Finally, a candidate constraints set is established to accelerate consensus by expanding the scope of constraints while preserving connectivity. In this study, connectivity and convergence of the system are theoretically analyzed from a geometric perspective. Simulation experiments are conducted to compare the proposed method with existing approaches under different densities and topologies. Simulation results demonstrate the superiority of this method in achieving fast convergence, particularly in large-scale and high-density scenarios.