1: \begin{abstract}

2:   The \emph{push-sum} algorithm allows distributed computing of the

3:   average on a directed graph, and is particularly relevant when one is restricted to one-way and/or asynchronous communications. We investigate its behavior in the presence of unreliable communication channels where messages can be lost.

4: %   in which case standard consensus algorithm generically fail.

5: %     communications are asynchronous  the context of directed and/or asynchronous communications} This is a useful method as directed

6: %  communication prevents standard consensus algorithms to compute the

7: %  exact average in general. 

8: %  In this paper this algorithm is investigated in the presence of an

9: %  unreliable communication channel where messages might be lost. 

10:   

11:   We show that convergence still holds, and analyze the error of

12:   the final common value we get for the essential case of two nodes,

13:   both theoretically and numerically. We compare this error

14:   performance with that of the standard

15:   consensus algorithm. For the multi-node case, we deduce fundamental

16:   properties that implicitly describe the distribution of the final

17:   value obtained.

18:  \end{abstract}

19: