Improving short-term information spreading efficiency in scale-free networks by specifying top large-degree vertices as the initial spreaders

Abstract

The positive function of initially influential vertices could be exploited to improve spreading efficiency for short-term spreading in scale-free networks. However, the selection of initial spreaders depends on the specific scenes. The selection of initial spreaders needs to offer low complexity and low power consumption for short-term spreading. In this paper, we propose a selection strategy for efficiently spreading information by specifying a set of top large-degree vertices as the initially informed vertices. The essential idea behind the proposed selection strategy is to exploit the significant diffusion of the top large-degree vertices at the beginning of spreading. To evaluate the positive impact of initially influential vertices, we first build an information spreading model in the Barabási–Albert (BA) scale-free network; next, we design 54 comparative Monte Carlo experiments based on a benchmark strategy and the proposed selection strategy in different BA scale-free network structures. Experimental results indicate that (i) the proposed selection strategy can significantly improve spreading efficiency in the short-term spreading and (ii) both network size and number of hubs have a strong impact on spreading efficiency, while the number of initially informed vertices has a weak impact. The proposed selection strategy can be employed in short-term spreading, such as sending warnings or crisis information spreading or information spreading in emergency training or realistic emergency scenes.

An Efficient Spreading Strategy Considering Information Decays and Partial Interactions Between People in Scale-Free Networks

“…Moreover, the configurations of the proposed model have been introduced in Section II, and except the spreading strategies, those configurations are constant in our simulations to avoid the impact of other factors on the spreading efficiency. The strategy that large-degree vertices are specified as the initial spreaders [21] is the benchmark strategy in our experiments.…”

“…Liu et al [19] and Pei and Makse [20] indicated that the spreading efficiency would be improved when vertices preferentially spread the information to large-degree neighbors. Wang et al [21] demonstrated that the spreading is more efficient when the large-degree vertices are specified as the initial spreaders. However, the important impact of smalldegree vertices on the spreading efficiency is gradually discovered by people.…”