A computation offloading scheme based on age of information for substation power IoT

Abstract

With the continuous development of power internet of things (PIoT), many smart IoT devices, such as drones, robots, and infrared cameras, are now widely used in substation inspections, resulting in a large amount of IoT task data. Furthermore, additional support is necessary to facilitate data processing for most computation-intensive applications due to limitations in computing resources and the endurance of smart devices (SDs). Therefore, this paper proposes an age of information (AoI) based computing task offloading scheme (ATO) for substation PIoT to solve the data processing problem. This paper utilizes time-average AoI to measure the freshness of the task computation results received by the user platform and develops a freshness model of task data processing for substation inspection scenario. We construct a system cost function by considering both the system’s time-average AoI and the energy consumption of wireless smart devices (WLSDs). We formulate a constrained optimization problem to minimize the system cost. Additionally, we propose a nonlinear parameter-improved Grey Wolf optimizer (ATO-NGWO) for the proposed ATO problem. The simulation results demonstrate that the ATO-NGWO can reduce average system cost by 2.4%, 6.5%, 9.5%, and 7.4% compared to ATO Grey Wolf optimizer (ATO-GWO), ATO Genetic Algorithm (ATO-GA), ATO Particle Swarm Optimization (ATO-PSO), and ATO Random Selection Algorithm (ATO-RSA). In cases where the computing capacity of the edge service center (ESC) is low, ATO-NGWO can lower system cost by 24% and 30% compared to All-Local Computing (ALC) and All-Edge Computing (AEC), respectively. In cases where ESC’s computing capacity is sufficient, ATO-NGWO can reduce system cost by 34% and 11.2%.