Comparative Study of Distributed Consensus Gossip Algorithms for Network Size Estimation in Multi-Agent Systems

Determining the network size is a critical process in numerous areas (e.g., computer science, logistic, epidemiology, social networking services, mathematical modeling, demography, etc.). However, many modern real-world systems are so extensive that measuring their poses serious challenge. Therefore, algorithms for determining/estimating this parameter an effective manner have been gaining popularity over past decades. In paper, we analyze five frequently applied distributed consensus gossip-based estimation multi-agent (namely, Randomized gossip algorithm, Geographic Broadcast Push-Sum protocol, and Push-Pull protocol). We examine performance of mentioned with bounded execution random geometric graphs by applying two metrics: number sent messages required achievement precision quantified as median deviation from real value size. The experimental part consists scenarios—the conditioned either values inner states or estimates—and, both scenarios, best-connected worst-connected agent chosen leader. goal paper to identify whether all examined applicable estimating size, which algorithm provides best performance, how leader selection can affect algorithms, most effectively configure stopping criterion.