Research On Edge Computing Task Offloading And Application Allocation Strategy In IoV

In recent years,with the development of Internet of Vehicles(IoV),computationally intensive on-board applications such as autonomous driving and high-precision 3D mapping have emerged,whose tasks are quite sensitive to latency and consume large amounts of computing resources.Due to the hardware cost and energy consumption,it is difficult for the on-board computing unit to meet the increasing demand for computing resources.Consequently,offloading the computing task to other platforms has become a mainstream research direction.Offloading the on-board application tasks to the edge servers in Mobile Edge Computing(MEC)is considered as one of the promising solutions to improve the quality of on-board application services.With abundant computing resources,storage resources and communication resources,edge servers provide a low-latency and reliable computing environment for IoV,thus relieving the computing pressure of vehicles and improving the data transmission efficiency of IoV system.This thesis studies the task offloading strategies and application placement algorithms in edge computing environment of vehicular networks.It first divides the task offloading process into two stages: offloading decision and task executing,and optimizes them respectively.Then,it proposes a clustering edge server application placement algorithm to improve resource utilization efficiency and reduce task offloading latency.The main works of the thesis are as follows:1.In view of the current insufficient vehicle computing capability and diversified vehicle tasks,this thesis studies the offloading strategy in the offloading scene of vehicular network edge computing,aiming to minimize the delay,and proposes a cloud-edge collaborative offloading strategy optimization(CEOS).In the offloading decision stage,the cloud-edge collaborative offloading strategy is proposed to accelerate the decision speed and reduce the computing pressure of vehicles,making full use of the computing resources in the vehicular network.In the task execution stage,a cache system is applied to the execution process to solve the problem of resource waste caused by repeated calculations,and to alleviate the low cache hit rate of computing tasks,a task decomposition model is constructed to decompose the task into several subtasks.Subsequently,considering the multiple factors affecting the cache value of the task result,the Multi-factor Priority(MFP)Cache Replacement Algorithm is proposed.Finally,the simulation experiment shows that compared with the traditional strategy,the proposed strategy can effectively reduce the completion delay of computing tasks and improve the system resource utilization rate.2.Considering the high coupling between tasks and applications services in the edge offloading scenarios,as well as the limited storage space of edge servers,this thesis further proposes a clustering edge server placement strategy to provide edge offloading services to vehicles,with the nodes in the cluster cooperating with each other to give the cluster the ability to process diversified tasks.On this basis,in order to reduce the extra system cost caused by the misalignment of tasks and applications in the cluster,this study further proposes an Adaptive Gradient Particle Swarm Optimization(AGPSO)to obtain the corresponding placement relationship between the on-board applications and the edge servers,and optimizes the on-board application installation process through Docker virtual container technology.Simulation results show that compared with the comparison algorithm,the proposed algorithm can effectively improve the overall offloading success rate and reduce the offloading delay.