Research On Task And Resource Optimal Scheduling Algorithms For MEC Enabled 5G Internet Of Vehicles

With the increasing number of cars,the development of the Internet of Vehicles has attracted much attention.Advances in Io T technology in terms of environmental perception,communication,computing,etc.have also promoted the further large-scale application of the Internet of Vehicles,but the existing network architecture of the Internet of Vehicles still cannot meet the low-latency business requirements of fast-moving vehicles.The giant connection and ultra-low latency characteristics of 5G mobile communication can further optimize the solution to the communication problem of the Internet of Vehicles.Although vehicle computing equipment is being updated,the vehicle’s computing resources are still unable to cope with high-precision maps,media entertainment and other Internet of Vehicle applications,and machine learning-based services such as autonomous driving require extremely strong computing power.MEC(Multi-Access Edge Computing)generally deploys computing resources on the side of the base station,and the application can provide services for users at a short distance,which not only meets the computing power requirements,but also greatly reduces the communication cost.The time delay avoids the problems of long delay and uncertain delay when accessing the remote centralized cloud.However,under the background of numerous computing tasks and massive data in the Internet of Vehicles,MEC is still a huge challenge to enable task scheduling and resource allocation of 5G Internet of Vehicles.Therefore,in the context of MEC enabling 5G Internet of Vehicles,this thesis studies the problem of how to select offloading nodes and allocate computing resources online for independent tasks and dependent tasks.The main content of this article is summarized as follows:(1)Aiming at the task scheduling problem after the vehicle offloads independent tasks to the MEC in the MEC-enabled 5G Internet of Vehicles scenario,this thesis proposes a task scheduling algorithm LBPM based on Lyapunov optimization theory for MEC collaborative computing to maximize the MEC provider Time average profit.LBPM proposes a new Lyapunov queue establishment method,and then establishes the Lyapunov function to model all variable-length tasks offloaded to the MEC as multi-queue task scheduling,and then make a reasonable unloading decision in each time slot to solve the task allocation problem of heterogeneous multi-server.In this thesis,the LBPM algorithm is compared with the existing algorithm,and the results show that the LBPM algorithm increases the time-average profit by more than 20%.(2)Aiming at the subtask scheduling problem after the vehicle initiates a single dependent task in the MEC-enabled 5G Internet of Vehicles scenario,this thesis proposes a subtask scheduling algorithm PBTSA based on priority greedy offloading to minimize task processing delay.PBTSA proposes a method that can better measure the data transmission and calculation delay of the Internet of Vehicles network,and then uses the RBFS(Reverse Breadth-First Search)algorithm to generate the priority of each subtask,and finally subtasks are greedily offloaded with low complexity according to priority to minimize task processing delay.In this thesis,the PBTSA algorithm is compared with the existing algorithms,and the results show that the PBTSA algorithm can effectively reduce the task processing delay,and the decision algorithm running time also has advantages.