Research On The Offload Strategy Of MEC-Based IoV

Many in-vehicle services consume a lot of computing resources and have strict delay requirements.The computing capabilities of existing in-vehicle terminals cannot meet large-scale computing needs.Mobile Edge Computing(MEC)provides the possibility to solve such problems.It pushes computing resources to the access network and provides computing services near the vehicle.In this thesis,based on the fast-moving Internet of Vehicles scenario,from the perspective of reducing the task completion delay and prioritizing on-board safety computing tasks,three total offloading strategies based on partial and binary offloading are designed.When the vehicle speed is too fast,the vehicle may cross the coverage of multiple Road Side Units(RSU)within the time limit of a single computing task.The calculation result is often not the server that handles the calculation task,but requires multi-hop communication to use other servers to achieve data interaction.When the number of tasks is large or the amount of calculation is large,within the time of completion of the calculation task offloaded to the MEC server,the vehicle drives away from the RSU within the current range.To address this problem,this thesis proposes a joint V2 V partial offloading strategy(JP-OS)and a MEC status-based partial offloading strategy(MSPOS).The JP-OS strategy combines the two offloading paths of V2 V and V2 I offloading to find the best splitting factor to split the task into two parts.The MSP-OS strategy performs selective offloading according to the load status of the MEC server.By analyzing the relationship between the calculation delay and transmission delay of the calculation task,a better task division ratio is determined,and the calculation task is divided into multiple MEC servers.Obtain more computing resources for the task and reduce the task calculation delay.Finally,the experimental simulation proves that the two calculation offload strategies can save more task delay when calculating the task data volume and calculation volume.From the perspective of giving priority to on-board safe computing tasks,this thesis designs a genetic algorithm based offloading strategy(GA-OS).A one-way onedimensional road V2 I offload model was established to analyze the movement of vehiclemounted terminals,analyze the computing resources consumed by tasks,and prioritize tasks for tasks.A genetic algorithm is used to encode the computing tasks,and based on the adaptive function of task weights,the tasks are offloaded to each server.The simulation verifies that the GA-OS strategy can achieve priority processing of vehiclemounted safety computing tasks compared with the traditional offload strategy,and increase the overall proportion of successful task processing,so as to realize the calculation of the MEC server in the case of uneven distribution of computing resources.The task is effectively uninstalled.