Performance research on a task offloading strategy in a two-tier edge structure-based MEC system

With the rapid development for the technology of Mobile edge computing (MEC), tasks tend to be more diversified and personalized, but fewer scholars considered differentiated quality of service requirements from diverse tasks in task offloading studies. In order to guarantee the real-time performance of latency-sensitive tasks and the throughput of latency-tolerant tasks, we propose a task offloading strategy in a MEC system with a two-tier edge structure. We establish a system model composed of a local model and an edge model to capture the workflow of tasks based on our proposed task offloading strategy, and we derive the key performance measures in terms of the average delay of a latency-sensitive task, the average delay of a latency-tolerant task, the utility of MBS Cluster I and the average power of the MEC system. We carry out experiments with analysis and simulation out to evaluate the long-term performance and validate the effectiveness of our proposed task offloading strategy. Finally, by trading off the average delay of a task and the average power of the MEC system, we formulate an optimization problem with inequality constraints for the average delay of a latency-sensitive task and the stability conditions of the MEC system. Furthermore, we develop an improved Powell–Hestenes–Rockafellar algorithm based on Lagrangian multiplier method to jointly optimize the offloading probabilities of the two types of tasks.