Accelerating the convergence rate of distributed consensus algorithm based on label propagation algorithm

Abstract

In order to accelerate the convergence rate of distributed consensus problem under complex topology, in the paper, the distributed consensus algorithm based label propagation algorithm was proposed. Firstly, we composed the complex topology into two layer of topologies by label propagation algorithm, the first layer of topology was consist of a few small communities, and every small community was considered as a node of the second layer of topology. The consensus firstly was reached in the first layer topology, then the consensus was reached in the second layer topology. In the paper, the convergence performance of the algorithm was proved. The analysis and simulation on convergence rate were done, the results show the convergence rate of the algorithm was higher than that of the usually first-order distributed consensus algorithm.