1: \begin{abstract}

2: By the distributed averaging problem is meant the problem of computing the average value of a

3: set of numbers possessed by the agents  in a distributed network using only  communication between neighboring agents.

4: Gossiping is a well-known approach to the  problem which seeks to iteratively arrive at a solution  by

5: allowing each agent to interchange information with at most one neighbor  at each iterative step. Crafting a

6: gossiping protocol which accomplishes

7: this is challenging because gossiping is an inherently collaborative process which

8: can lead to deadlocks unless careful precautions are taken to ensure that it does not.

9: Many gossiping protocols are request-based which means simply

10: that a gossip between two agents will occur whenever one of

11: the two agents accepts a request to gossip placed by the other.

12: In this paper, we present three deterministic request-based protocols.

13: We show by example that the first can deadlock.

14: The second

15: is guaranteed to avoid deadlocks and  requires fewer transmissions per iteration than

16: standard broadcast-based distributed averaging protocols

17: by exploiting the idea of local ordering together with the notion of an agent's neighbor queue;

18: the protocol requires

19: the simplest queue updates, which provides an in-depth understanding of how local

20: ordering and queue updates avoid deadlocks.

21:  It is shown that a third protocol

22: which uses a slightly more complicated queue update rule can lead to significantly

23: faster convergence; a worst case bound on convergence rate is provided.

24: %An asynchronous version of the third protocol is also provided

25: %which is shown to solve the distributed averaging problem in the case when

26: %each agent has its own clock.

27: \end{abstract}