Distributed average consensus via gossip algorithm with real-valued and quantized data for

This paper studies the problem of the gossip consensus algorithm with real-valued and quantized data. We study the effect of the mixing parameter on the convergence rate of the proposed gossip consensus algorithm, and show when the proposed bounds are optimized with respect to the mixing parameter. For a gossip consensus algorithm with quantized data, we prove that it can achieve the consensus almost surely, and the expected value of the final states is equal to the average of the initial states. Moreover, we provide a result characterizing the convergence performance of the distance from consensus and make a comparison with the non-quantized gossip consensus algorithm. Finally, simulation results are provided to evaluate the effectiveness of the proposed algorithm.