1: \begin{abstract}

2: In this paper we analyse Belief Propagation over a Gaussian model in a dynamic environment.

3: Recently, this has been proposed as a method to average local measurement values by a distributed 

4: protocol (``Consensus Propagation'', Moallemi \& Van Roy, 2006), where the average is available 

5: for read-out at every single node. In the case that the underlying network is constant but the

6: values to be averaged fluctuate (``dynamic data''), convergence and accuracy are 

7: determined by the spectral properties of an associated Ruelle-Perron-Frobenius operator.

8: For Gaussian models on Erd\H{o}s-R\'{e}nyi graphs, numerical computation

9: points to a spectral gap remaining in the large-size limit, implying exceptionally 

10: good scalability. In a model where the underlying network 

11: also fluctuates (``dynamic network''), averaging is more effective than in

12: the dynamic data case. Altogether, this implies very good performance of these

13: methods in very large systems, and opens a new field of statistical physics

14: of large (and dynamic) information systems.  

15: \end{abstract}

16: