Research On Cloud Resource Management And Optimization Under The Application Of Internet Of Vehicles

With the strong development of Intelligent Transportation System(ITS),intelligent vehicles are running a large number of computationally intensive applications,which generate a large amount of task data that needs to be processed quickly and in real time.In order to obtain a high-quality service experience,vehicles are required to have powerful computing and storage capabilities.However,in reality,the resources of vehicles are limited.The Internet of Vehicles combined with cloud computing technology has become the inevitable trend to solve the above problems.In this paper,under the three-layer cloud network architecture of "remote cloud-edge cloud-vehicle cloud",the research on the optimization of cloud resource management under vehicle networking applications was carried out.The main content and work results were summarized as follows:(1)Under the three-layer cloud network architecture,aiming at the problem of limited computing resources of vehicles in the Internet of Vehicles and the high cost and high delay time of traditional cloud computing resources,a Multi-cloud Collaboratively assisting Vehicle Computing(MCAVC)paradigm was defined.A joint optimization scheme for offloading node selection and resource allocation based on improved Genetic Algorithm(GA)was proposed,which minimized the weighted sum of the task completion time and the cost of computing resources(total computational overhead)by improving coding,crossover and mutation operations.The simulation results show that the MCAVC scheme has a smaller total computational cost compared with the other three existing schemes.Compared with the other two algorithms,the improved algorithm has lower convergence value in total computational cost,and its performance is better.(2)From the perspective of the lowest layer of the hierarchical cloud network architecture--vehicle cloud,aiming at the problem of the limited computing/storage resources of vehicles in the Internet of Vehicles and the characteristics of vehicle mobility,a migration mechanism of "cross area cloud service" was proposed,and the service migration decision and dynamic resource allocation of vehicle cloud system based on Semi-Markov Decision Processes(SMDP)were mainly studied.Different types of vehicles were considered in the vehicle cloud system,and finally the value iterative algorithm was applied to solve the optimal strategy of resource allocation.The simulation results prove that different types of vehicles have a certain impact on resource allocation.At the same time,SMDP resource management scheme has a lower request rejection rate than Random Allocation(RA)strategy and Greedy Algorithm(GA),and the long-term benefits of the system is increased by 3.8% and10% respectively,which proves the superiority of SMDP resource management scheme.