A Correction to Algorithm A2 in “Asynchronous Distributed Averaging on Communication Networks”

This paper discusses Algorithm A2 in “Asynchronous Distributed Averaging on Communication Networks” (IEEE Trans. Netw., vol. 15, no. 3, pp. 512-520, Jun. 2007), which claims to solve the distributed averaging problem provided that the parameters to the algorithm meet certain constraints. Specifically, the states of each node in the network are claimed to converge to the average of the initial values associated with the nodes under these constraints. This paper points out a flaw in the proof of the algorithm and in addition provides a specific example of a network, satisfying the assumptions, for which the algorithm does not converge.