Research On Task Offloading Strategy Of Vehicular Edge Computing In Urban Environment

The rapid development and application of 5G communication technology has accelerated the informatization and intelligentization of the whole society.Vehicular Adhoc Networks(VANETs),an important part of the Intelligent Traffic System(ITS),has also evolved with it,showing that vehicles are becoming more and more informatized,networked,and intelligent.In this process,more and more computing-intensive vehicle applications have emerged,posing severe challenges to the limited computing resources of the vehicles themselves.To this end,Mobile Edge Computing(MEC)technology is introduced into the VANETs to offload the complex computing tasks of the vehicle to the MEC server for execution,thereby reducing the computing pressure of the vehicle.The new computing model generated by the combination of vehicle and mobile edge computing is called vehicular edge computing(VEC),and how to achieve effective offloading decisions and resource allocation of computing tasks on the premise of meeting the low latency requirements of vehicle computing tasks,is a key issue to be solved urgently in the field of VEC research.The dissertation aims to improve the task completion rate of vehicles in VANETs and reduce the time required for task completion,and conducts research on two vehicular edge computing scenarios in an urban environment.The main research work is as follows:(1)For scenarios where there are multiple MEC servers,multiple vehicles and each vehicle has multiple independent computing tasks,consider the task offloading and local execution methods,the problem of multi-task offloading in multiple MEC servers is studied.A vehicular task offloading algorithm based on Q-learning is proposed to solve the problem.In this algorithm,the reward function in Q-learning is reconstructed,and factors such as task completion time are introduced.The simulation results show that the proposed task offloading algorithm based on Q-learning has less average task completion time than other comparison algorithms.(2)For scenarios where there are multiple MEC servers,multiple vehicles,and each vehicle has multiple dependent computing tasks,a directed acyclic graph is used to describe the dependencies between tasks,and the offloading decision-making of dependent tasks and the selection of MEC servers are constructed as an optimization model for problem solving.Under the premise of comprehensively considering factors such as vehicle mobility,server collaboration,and multiple execution methods,a prioritybased heuristic algorithm is designed to solve this optimize problem.Solve.The simulation results show that the proposed heuristic algorithm has a significant improvement in application completion rate and average application completion time indicators compared with existing schemes.In summary,aiming at the problem of task offloading of vehicle edge computing in urban environments,the dissertation separately studied two typical task offloading strategies in vehicle edge computing scenarios based on meeting the requirements of vehicle mobility and task delay.The research work of the dissertation and the research results obtained will not only help to improve the intelligence and information service level of vehicles,but also help to provide solid technical support for the construction of intelligent transportation systems,which has important theoretical and practical significance.