Workload Allocation Based on User Mobility in Mobile Edge Computing

Abstract

Mobile Edge Computing (MEC) has become the most possible network architecture to realize the vision of interconnection of all things. By offloading compute-intensive or latency-sensitive applications to nearby small cell base stations (sBSs), the execution latency and device power consumption can be reduced on resource-constrained mobile devices. However, computation delay of Mobile Edge Network (MEN) tasks are neglected while the unloading decision-making is studied in depth. In this paper, we propose a workload allocation scheme which combines the task allocation optimization of mobile edge network with the actual user behavior activities to predict the task allocation of single user. We obtain the next possible location through the user's past location information, and receive the next access server according to the grid matrix. Furthermore, the next time task sequence is calculated on the base of the historical time task sequence, and the server is chosen to preload the task. In the experiments, the results demonstrate a high accuracy of our proposed model.