With the rapid improvement of software service in recent years, compute-intensive services such as VR, AR, and face recognition have been developed drastically. It's hard for mobile devices with limited capacity computing and energy to meet the delay requirements of these services. With the cloud computing technology, the high computing demands for these compute-intensive services can be satisfied by migrating these compute-intensive services to powerful computation servers in the cloud. However, it will bring extra latency by long-distance communication between servers and mobile devices. Edge computing is emerging to reduces the extra latency coursed by sinking services to the edge. In this paper, we proposed an intelligent service migration algorithm model based on machine learning algorithms. Considering the dynamic of the bandwidth of network and the battery level of mobile devices, the proposed algorithm is aimed at optimizing the delay and energy consumption. The simulation shows that the performance the proposed service migration algorithm is better than the comparative algorithms.
[1]
Xing Chen,et al.
AndroidOff: Offloading android application based on cost estimation
,
2019,
J. Syst. Softw..
[2]
Rui Zhang,et al.
Accelerated multiscale space–time finite element simulation and application to high cycle fatigue life prediction
,
2016
.
[3]
Shane Legg,et al.
Human-level control through deep reinforcement learning
,
2015,
Nature.
[4]
Ahmed Ghoneim,et al.
Intelligent task prediction and computation offloading based on mobile-edge cloud computing
,
2020,
Future Gener. Comput. Syst..
[5]
Yoshiaki Tanaka,et al.
A Deep Reinforcement Learning Based Approach for Cost- and Energy-Aware Multi-Flow Mobile Data Offloading
,
2018,
IEICE Trans. Commun..