Research On MEC-Based Task Offloading And Migration Mechanism In Internet Of Vehicles

With the development of the Internet of Vehicles,various vehicle applications require lower task processing latency.The computing capability of vehicles is limited,so Mobile Edge Computing(MEC)technology is introduced into the Internet of Vehicles.Cloud services are transferred to the edge of the wireless access network,and users can offload computing-intensive tasks to the edge of the network in order to relieve the pressure of local computing.In addition,to satisfy the needs of emerging applications,the vehicular network architecture must provide more serveices with flexibility,programmability,reliability,and scalability.Thus,MEC is combined with the Software Defined Network(SDN)architecture to improve the system performance.However,in the MEC-based softwaredefined vehicular network,there still exists the problem of high task processing delay.Therefore,this thesis discusses the user task offloading and migration delay optimization problem.The main contributions are presented as follows:1.The delay optimization problem of task migration of software-defined vehicular network based on MEC is studied.Due to the movement of vehicles,the undone computing tasks need to be migrated from the original roadside unit to the target roadside unit,leading to the migration problem of virtual machines.Model the task migration delay optimization problem as a mixed integer nonlinear problem,and the original optimization problem is divided into two sub-optimization problems through two-stage algorithm.Utilize the deep programmability of SDN to reconfigure virtual machine migration routes dynamically.The Dijkstra routing algorithm is employed to find the best path for this task migration.Then,the Q-learning algorithm is adopted to optimize the migration delay of the current task,which consequently reduces the total task delay when the vehicle users are moving and enhances the user service experience.The simulation results demonstrate that this scheme can effectively lower the user’s service delay.2.The optimization of task offloading,caching and migration delay of the Internet of Vehicles based on NOMA-MEC is studied.In the Internet of Vehicles that combining MEC and NOMA technology,addressing at the high latency problems faced by users when processing computationally intensive and latency-sensitive tasks,an optimization strategy of task offloading,migration and cache based on game theory and Q learning is proposed.First,the offloading delay,migration delay and cache delay of the Internet of Vehicles task are constructed based on NOMA-MEC.Then,the cooperative game model is applied to obtain the optimal user grouping to optimize the offloading delay.Finally,in order to avoid local optima,the Q learning algorithm is involved to optimize the joint delay of the migration cache in the user group.The simulation results indicate that the proposed scheme can effectively improve the offloading efficiency and reduce the task processing delay.