Reschedulable Task Allocation Strategy in Cloud-Edge-End Cooperative Mobile Crowd Sensing

In centralized mobile crowd sensing (MCS), the cloud platform assigns all the tasks to participants every time. Since the cloud platform consumes a lot of computing and communication resources to provide services for participants, it will bring about high communication delay and request congestion. The cloud-edge-end architecture for service provisioning has aroused extensive attention recently, owing to its advantages in resource provisioning in close proximity to the resource requestors. Despite the advantages of this architecture, we also observe that it cannot dynamically adjust the allocation scheme when the corresponding computing services are not available to the participants after the initial task allocation. To address this issue, we put forward a re-schedulable task allocation approach in the cloud-edge-end architecture. We aim to improve the efficiency of task execution such as the maximization of task completion rate, while considering service types provided by edge servers and multiple constraints such as resource balancing on the edge servers and deadlines for the task responses. An improved Grey Wolf Optimization (GWO) algorithm is adopted for task rescheduling in this paper. Simulation results indicate that the proposed algorithm performs well in terms of task completion rates and task average response time.