改进的分布式consensus在小世界网络中的应用

Modified Distributed Consensus in Small World Network

研究了一种改进的分布式consensus及其在小世界网络中的应用。通过将RH问题降到一维且使用惩罚函数F(x,x(i))来达到consensus。只要改变很少的几个连接(即改变p的大小),就可剧烈的改变网络的性能,从而加速改变收敛速度。通过代数连通度2λ(G)增大可提高一致收敛的速度。仿真结果表明该方法不但减少了主体之间的通信量,而且加快了一致性收敛的速度。

The distributed consensus algorithm and its application in small world network was investigated, with reduce the RH problem to 1-dimension and trace the fast signal to reach consensus with the penalty function F（x,x^（i））. The convergence speed of the consensus algorithm on a regular lattice can be greatly enhanced by rewiring a very small number of links in the network. Increasing the algebraic connectivity λ2 （G） and the Laplacian matrix λ2 （L）, could improve the convergent speed. The simulation results show that the scheme could reduce the communication among agents, and improve the convergent speed at the same time.