Research On Computing Offloading Method Based On MEC Orient To Internet Of Vehicle

The development of Io T and wireless technologies has paved the way for new applications with advanced capabilities in vehicles.For example,in-vehicle cameras and embedded sensors can play a key role in efficient and safe transportation systems.However,resource-constrained vehicles can be constrained by computationally intensive applications,causing bottlenecks that make it difficult for vehicles to ensure the required quality of service（Qo S）levels.Mobile edge computing,as a new architecture and key technology for emerging 5G networks,has been proposed to solve this problem.However,considering that the resources of MEC are not unlimited,the continuous increase in the number of offload tasks will cause the load of the MEC server to exceed its maximum limit,which makes MEC unable to guarantee Qos for the application of each vehicle,in which case the vehicle will not benefit from computational offloading.Therefore,a reasonable offloading strategy and resource allocation will help improve the overall performance of the network,and will ensure the Qo S of vehicle users.Faced with these problems,this paper has done the following research:（1）For the limitation of MEC computing resources,in the face of divisible computing tasks,partial offloading technology is adopted.Considering that the offloading ratio and resource allocation will affect each other,they are jointly optimized,and an iterative optimization scheme based on block coordinate descent technology,joint convex optimization and Sparrow Algorithm Fusion Difference and Cauchy Variation（BCD-CONISSA）is proposed.The convex optimization algorithm is used to solve the resource allocation problem.Since the sparrow search algorithm has the characteristics of high solution accuracy and strong robustness,it is used to solve the offloading ratio problem.In order to avoid the reduction of population diversity in the later stage,the basic sparrow algorithm is combined with Difference and Cauchy mutation.Then,based on the block coordinate descent technique,the two sub-problems of resource allocation and offloading ratio are iteratively optimized and solved.Simulation experiments show that the BCD-CONISSA offloading algorithm proposed in this paper reduces the latency,energy consumption and cost of computing tasks compared to the benchmark scheme.（2）For the multi-vehicle and multi-server scenario,in the face of inseparable computing tasks,the complete offloading technology is adopted.Multiple tasks are burst to the server at the same time.The offloading strategy will affect the performance gain of the entire network.Therefore,the user’s offloading strategy and the MEC calculation are combined.Resource allocation is jointly optimized,and a two-stage heuristic algorithm is proposed.In the process of modeling,the optimization of latency and energy consumption is transformed into optimization of cost by introducing weight factors.The optimization goal is to reduce the average cost in the system.Since the problem to be solved is a typical mixed integer nonlinear programming problem,which is usually NP-hard,the problem is divided into two sub-problems.In the first stage,the algorithm uses an improved hybrid genetic algorithm to solve the offloading strategy problem.The greedy correction method is applied to the chromosomes that do not meet the constraints and the chromosomes that meet the constraints but have not fully utilized MEC computing resources to convert them into chromosomes that fully utilize resources,thereby improving the convergence speed and optimization ability of the algorithm.In the second stage,the improved artificial fish swarm algorithm is used to solve the problem of resource allocation.The basic artificial fish swarm algorithm step size and field of view are generally fixed values,and the early search is easy to fall into blind search.Therefore,an adaptive factor is introduced to make the field of view and step size dynamically adjust according to the changes of external environment information,so that the algorithm has better global search ability in the early stage,and can also avoid the phenomenon of oscillation that is easy to occur in the later stage of the algorithm.The simulation results show that,in the face of multi-vehicle and multi-server scenarios,this scheme can generate lower latency,energy consumption and cost than the benchmark scheme.