Research On Dynamic Computation Offloading Algorithm In Vehicular Edge Computing Environment

Mobile edge computing technology is a promising mobile computing mode,which can meet different application requirements such as low latency and high reliability in 5G scenarios.In the edge computing scenario,users can offload complex computing tasks from the mobile devices that cannot be completed to idle high-performance devices to perform computing in different communication ways to reduce the response time of tasks.In the future,the system integrated with the Internet of Vehicles,V2X(Vehicle to Everything)and mobile edge computing can provide more accurate and comprehensive perception information for vehicles through sensor information sharing,vehicle road cooperation and other technologies.However,due to the high-speed mobility and dynamic resources of the Internet of vehicles,the static computation offloading algorithm can no longer meet the needs of mobile computing task.This paper research on how to reduce the impact of environment and resource changes on task execution,and the main work is as follows: Firstly,this paper considers that there are many idle vehicles with abundant resources in the road,which can be used as mobile servers to perform computing tasks,enriching the choice of users’ computation offloading decision.In the Vehicle to Vehicle(V2V)offloading environment,a partial flooding computation offloading algorithm based on link reliability and computing power is proposed.The offloading nodes are selected according to the communication reliability and computing power of nodes.The system reliability and resource utilization are balanced while the task completion delay is reduced.Simulation results show that the proposed algorithm can select an appropriate number of better agent resource nodes to perform the computing,which combines the advantages of traditional algorithms.It can not only ensure the low delay requirement of tasks,but also improve the utilization of network resources.Then,to deal with the dynamic of resources in the Cellular-V2X(C-V2X)vehicle edge computing environment,this paper proposes an Environment Aware Dynamic Migration Strategy based on Hidden Markov Model(EADM).In this paper,HMM model is used to describe the change process of server performance,and the computing performance of server is predicted,which reduces the ping-pong effect caused by the change of environment information in the process of computation migration.In addition,EADM algorithm dynamically adjusts the current migration strategy according to the perceived changes in the environment,which reduces the impact of the mobility of Internet of vehicles users on the task execution process.The experimental data prove that the EADM algorithm can greatly cut down the migration cost caused by ping-pong effect,adapt to the dynamic changes of the Internet of vehicles,and meet the continuous needs of users with high performance and low delay requirements.