Index Geographic Gossip Algorithm for Information Dissemination over Regular Graphs

Abstract

An Index geographic gossip (IGG) algorithm is proposed. Relay nodes participate in information exchange and updating. The cumulative number of times these nodes participate is characterized by an index number, which can be used to accelerate information updating. The convergence property of the IGG algorithm is theoretically analyzed in ring and grid network topologies. The IGG algorithm improves the standard gossip algorithm by a gain of O(n) in both convergence time and communication cost. Compared to the geographic gossip algorithm, the IGG algorithm has a gain on the order of O(n) and O(n 1/2) in the average hop count for information exchange and communication cost, respectively. Finally, the proposed IGG algorithm is compared with various baselines through simulations, and it is shown that significant performance gain can be achieved.