Multi-user fine-grained task offloading scheduling strategy under cloud-edge-end collaboration

Abstract

To solve the current problems of inefficiency, low utilization of intensive network resources, and high system cost in handling multi-user applications, a multi-user fine-grained task offloading scheduling approach under cloud-edge-end collaboration was proposed. Latency, energy consumption, and server rental costs were jointly considered. Application tasks were firstly divided and subtask priorities were designed. Then, a multi-user subtask scheduling scheme was proposed and an improved simulated annealing particle swarm algorithm was designed to minimize the total system cost to achieve the optimal offloading decision. Experimental results show that the proposed method reduces the total cost by at least 12.28% and 7.42% compared to other methods such as particle swarm and simulated annealing binary particle swarm, respectively.

Keywords

• edge computing;task offloading;cloud-edge-end collaboration;multi-user;fine-grained scheduling