A distributed swarm intelligence‐based energy‐saving method among massive edge nodes

Abstract

Summary In edge computing, how to save energy among sustainable edge nodes is a hot topic. ON/OFF switching of edge nodes as a key point is efficient but still suffers from the long round‐trip time problem because of its centralized control manner. Especially in the wireless network, service coverage is proved to be NP‐Complete. To this end, we propose a Distributed Swarm intelligence‐based Energy‐saving algorithm (DSE). In DSE, pheromone and residual energy are used to calculate the wake‐up probability. Through the wake‐up probability, the edge node can be activated periodically and efficiently. In order to balance the energy in the whole system that contains massive edge nodes, we further use a correction factor, that is, DSE+, to adjust the wake‐up probability of the nodes. The proposed methods allow for distributed implementation without requiring a centralized control by the coordinator, and the pheromone accumulated temporally and spatially. In addition, they do not require node localization. Experiments show that both DSE and DSE+ can work as expected, and DSE+ with the correction factor improves the lifetime of the whole system at least 12.6% compared with the DSE without the correction factor.