Research On Edge Resource Allocation And Selection Algorithm Based On Vehicle Trajectory

With the continuous development of the social economy and the rapid development of the Internet of Things,vehicle applications and services are increasingly enriched.At the same time,users’ requirements for data rate and service quality are constantly improving,which is prone to a series of problems such as a large amount of data and low backhaul delay,and affect the service quality of the Io V.As a supplement and evolution of traditional cloud computing technology,edge computing technology can greatly reduce cloud computing load and data processing delay by deploying resources at the edge of the network close to connected vehicles,while providing a better solution for Io V applications.However,the edge computing technology of the Io V still faces some challenges.Too little or too much configuration of edge resources can easily lead to the problem of low utilization of edge resources.At the same time,the uneven load of different edge servers makes the success rate of requests for edge resources by connected vehicles lower.Therefore,designing efficient edge resource allocation and selection algorithms will help improve the quality of service(Qo S)of the overall Io V edge computing system.Aiming at the imbalance between the high cost of edge resource allocation and the low utilization of edge resources,this paper proposes a prediction and optimization framework for the allocation of a minimum number of edge resources.Firstly,according to the characteristics of the vehicle trajectory,the fine-grained model of the Kalman filter and the coarse-grained model of the Markov chain is used to predict the vehicle trajectory,and the fleet distribution vector and the fleet mobility matrix are created.Then,a probability distribution model is built to analyze the flow in and out of connected vehicles in a cell within a period,and the minimum number of edge resources is allocated to each cell under the premise that the blocking rate is not higher than the user’s predefined blocking rate.Finally,in order to improve the success rate of unloading the request of the connected vehicle to the edge resources,the optimal edge resource allocation algorithm is improved based on the characteristics of the summarized traffic flow combined with the model of traffic flow prediction.Experiments show that the scheme proposed in this paper is better than other edge resource allocation methods in weighing edge resource allocation cost and utilization.Aiming at the problem of load imbalance between different edge servers on the Internet of Vehicles edge system,queuing theory is used to propose the queuing computing model in the edge server,mobile vehicle terminal,and the small cloud node,and the allocation model of the requests from connected vehicles in this paper.First,based on the multivariate coupling problem in the optimal assignment problem,this paper divides the problem into two sub-processes: task assignment optimization and task redirection.Second,the task demands generated by connected vehicles are distributed among connected vehicle nodes,edge servers,and small cloud service areas using queuing theory and cost-increasing methods.Then,an iterative method and a transportation planning model are used to approximately solve the redirection flow between different cells,while calculating the minimum number of cloud servers to handle the remaining unprocessed task offloading requests.Finally,a dynamic selection algorithm for edge nodes by connected vehicles is proposed to determine the serial number of the edge server to be uninstalled or upload the task to the cloud node.Simulation experiments show that the proposed method of edge resource allocation and selection based on the prediction of vehicle trajectory can reduce resource allocation costs as much as possible while ensuring Qo S.