Edge-aided control dynamics for information diffusion in social Internet of Things

Abstract Information diffusion in social Internet of Things (SIoT) brings convenience to life, but it also faces challenges. The explosive growth of information exceeds the computing capability of the network, making it difficult to control. Motivated by the challenge, a novel control strategy for information diffusion is proposed in SIoT, which is based on the advantages of edge computing with low latency, high energy efficiency and timely feedback. Furthermore, graph neural networks (GNNs) is combined with ordinary differential equation systems (ODEs) to model edge-aided control dynamics, describing high-dimensionality and complicated interactions. Theoretical analysis and comparative simulation results show that the proposed edge-aided control strategy can dominate information diffusion within the target scope at the lowest cost.