Multi-relay Assisted Computation Offloading Strategy In Edge Computing Systems

With the rapid development of information and communication technologies,mobile devices can provide more and more content and service,which has been an important part of our daily life.Meanwhile,the demand for reliable and highperformance computing capabilities is also increasing.Therefore,Multi-access Edge Computing(MEC)is proposed to bring a better experience.In Multi-access Edge Computing with Energy Harvesting(MEC-EH),the mobile device can capture ambient recycling energy(e.g.,Solar,wind,and RF energy)and store them in a battery for further use.To execute a large number of computationally intensive tasks,as many computation-intensive tasks as possible should be offloaded to the MEC server.However,the wireless network is unstable and the movement of users unpredictable,which makes it difficult to have a stable MEC-EH system.When the system state is poor,for example,when the mobile device is at a long communication distance,poor communication link,or low energy level,it can greatly affect the quality of task execution.To address the problem,we consider the usage of D2 D communication and collaboration among users to improve the system performance.Based on the above analysis,the main work of this paper is as follows:First,we propose a single-relay assisted offloading strategy for the MEC-EH system.In addition to direct offloading and local execution,the mobile device is allowed to offload its tasks to the MEC server with the help of its neighbor node.Task execution cost is used as a performance matrix,which considers both task execution time and task drop rate.Furthermore,a Single-Relay Assisted Computation Offloading(SRACO)algorithm based on Lyapunov optimization is proposed.This algorithm includes the selection of the relay device and offloading strategy.Second,based on the single-relay assisted offloading strategy,we further consider the case that selecting multiple devices as relays.We propose a multi-relay assisted offloading framework in the MEC-EH system.Besides,we consider more complex scenarios in this work,such as the user mobility model.Further,the Multi-Relay Assisted Computation Offloading(MRACO)algorithm is proposed in this paper to minimize the average task execution cost.This algorithm includes multi-relay selection,computation task assignment,and offloading strategy selection.We evaluate proposed algorithms by using simulation experiments.The relayassisted computation offloading strategy effectively improves the performance of the system,especially when the system is in a poor state.The MRACO algorithm reduces about 22% compared to the SRACO algorithm and nearly 52% compared to the algorithm without the relay strategy,which proves the effectiveness of the proposed offloading strategy and algorithm.