Research On Mobile Edge Computing Task Offloading And Service Caching Under The Internet Of Vehicles

In the field of Internet of Vehicles,in order to meet users’ intelligent driving experience,various in car applications have entered people’s vision.Heterogeneous application services have different requirements for their execution and processing,and a large number of intelligent applications demand low latency,low energy consumption,and continuity of user experience for the network.When dealing with heterogeneous application services,simply reducing latency or energy consumption cannot bring a better application experience to all end users.At the same time,the system also needs to fully consider the working conditions of the terminal when initiating task requests.Therefore,it is necessary to allocate network resources and develop service strategies according to the current situation of different types of applications and terminals,in order to optimize system new capabilities and improve the quality of user experience(Qo E).This paper studies computing task offloading,edge computing resource allocation,service caching and task scheduling.The main research work includes the following two aspects:(1)In order to improve the resource shortage of edge computing(MEC)server and further improve the quality of user experience at the peak of user requests for services,the issues of task offloading,resource allocation and service caching of mobile terminals are jointly considered,and a collaborative caching and task scheduling method is designed.Firstly,the originally difficult and combinatorial problem is decomposed into a resource allocation subproblem with fixed offloading decisions and an uplink transmission power allocation subproblem.The fixed expression for computing resource allocation is obtained based on the KKT condition,and the optimal transmission power allocation is obtained using quasiconvex optimization;Secondly,based on the Zipf distribution to obtain the request probability model of application services,a cloud edge collaborative task active caching algorithm is designed for caching applications with high popularity in advance;Finally,based on the current power situation of the terminal,a task scheduling algorithm based on greedy strategy was designed to develop a service mode selection strategy for user tasks,in order to optimize user offloading efficiency.The simulation results show that the proposed caching algorithm and service mode selection algorithm can further reduce task processing latency and energy consumption,and improve the processing capacity of the MEC system.(2)Diversified and intelligent vehicle applications urgently need to reduce task execution latency and further ensure the continuity of user experience,while the vehicle’s own hardware configuration cannot meet these needs.How to efficiently use the limited edge computing resources to support the low latency,low energy consumption and high continuity required by on-board applications is the key technology to support the development of intelligent Internet of Vehicles.Based on the mobile edge computing system model of the Internet of Vehicles,and fully considering the vehicles with idle computing and storage capacity on the roadside,a vehicle communication link maintenance(VCIM)algorithm based on domain coverage is proposed.This algorithm selects the vehicle with the highest degree of communication retention within the cluster as the cluster head node.Other vehicles regularly transmit their own information,such as vehicle geographic location,driving direction,speed,and current CPU remaining computing resources,to the cluster head node.The cluster head node transmits the obtained vehicle information to the MEC server.When a vehicle sends a task request,MEC formulates a task offloading and resource allocation plan based on the vehicle information.The simulation results show that when the processing capacity of the MEC server reaches saturation,the cluster head node selects the optimal vehicle node for task processing based on the information sent by members within the cluster,effectively reducing task processing latency and energy consumption,and improving network resource utilization.