Fairness-oriented computation offloading for cloud-assisted edge computing

Abstract Mobile edge computing, which enhances the computing power of mobile devices via computation offloading technology, has been proposed as a promising paradigm to alleviate the mismatch between limited computational resources and ever-increasing computational requirements. An appropriate offloading strategy must be identified to optimize the computation offloading performance. Moreover, a mobile user is generally a selfish individual; therefore, we should consider the fairness problem when determining the offloading strategy. In this paper, we focus on computation offloading in a cloud-assisted edge computing system with multiple users, an edge server and a cloud server and aim to optimize the response time for each user. A fairness-oriented approach, including the determination of the application offloading strategy, the data transmission strategy and the cloud–edge cooperation strategy, is proposed. Initially, the computation offloading problem is formulated as a noncooperative game. Then, we find the optimal cloud–edge cooperation strategy and the optimal data transmission strategy based on a theoretical analysis and propose an iterative algorithm to identify the optimal offloading strategy for each user. Finally, we evaluate the proposed approach by means of extensive experiments, and the experimental results show that the proposed approach can significantly reduce the response time of mobile devices.