An Asynchronous Linear-Threshold Innovation Diffusion Model

Qi Zhang,Kuiying Deng,Tianguang Chu

DOI: https://doi.org/10.1007/978-981-10-2338-5_31

2016-01-01

Abstract:We consider a linear-threshold innovation diffusion model in asynchronous networks, wherein the positive vertices hold supporting attitude towards the innovation and accelerate the diffusion whereas the negative vertices oppose it and obstruct the diffusion. A more realistic assumption made in this work is that the attitudes of the vertices (i.e. positive or negative) are changeable. The diffusion begins from a small subset of active nodes and propagates towards the whole network through random transmitting delays along each edge. The vertices outside of the initial active set would be activated and keep forever in active status once the difference of the positive and negative signals they receive from their neighbors exceed a certain threshold. The attitudes of these active nodes would change if the difference of the received positive and negative signals reaches given thresholds. Under these assumptions, our model can exhibit an interesting phenomenon: whenever the initial active set surpasses a certain size, the diffusion process propagates to a specific proportion instead of all the vertices. Experiments on Erdős-Rényi networks verify that the theoretical estimations match closely with the experimental results.