Generalized Metropolis-Hastings Algorithm for Distributed Averaging with Uniform Quantization Scheme

Quantization allows mapping analog signals into a bit set of a finite size and is thus an essential process in real-life systems. In this paper, we analyze the generalized Metropolis-Hastings algorithm for an arithmetic mean estimation under a quantization noise. We vary the mixing parameter and the number of the bits for quantization in order to identify which initial configuration of the analyzed algorithm is the best performing and how the configuration of the applied uniform quantization scheme affects the estimation precision.