Research On Task Offloading Algorithm Based On Fog Computing In VANETs

With the rapid development of 5G technology,mobile applications such as autonomous driving,video streaming,and vehicle online games continue to emerge,and the data exchange and service requirements of portable terminal devices are also increasing.The rapid growth of data puts a heavy burden on the network and roadside unit(RSU),resulting in the inability of existing cellular networks to guarantee the quality of service for users.At the same time,the cost of maintaining network operations and setting up various equipment has also increased greatly,causing great difficulties for managers.In response to the above problems,this paper proposes a matching algorithm based on the kurh-munkras(KM)algorithm and a dynamic task offloading algorithm based on greedy,dedicated to reducing the processing delay of user tasks and improving the service experience of users in the vehicle environment.The main research contents of this paper are as follows:First,for the complex dynamic environment of vehicles in real life,a vehicle movement model(VMM)is established in the fog architecture environment.Then we define and process various parameters of different tasks and vehicles,and we turn the task offloading problem into a one-to-one matching problem.To solve this problem,a matching algorithm based on the KM algorithm(KMM)is proposed.Assuming that the task information is known,first,select the feasible offload server based on the relative distance between the vehicles and the communication constraints,then compare the calculation delay of task offloading to the feasible vehicle and the transmission delay between the two perform a second screening to obtain a suitable offload server,and finally,use The KM algorithm makes the final decision and assigns the corresponding offload server to each vehicle.Through one-to-one matching and uninstallation,the response time to complete the task is shortened and user satisfaction is improved.Secondly,considering the uncertainty of the vehicle environment(all the task information cannot be known at any time),this paper further proposes a greedy-based dynamic task offloading algorithm(GMDC).In a very short time,based on randomly arrived task requests,we can still find an offloading server suitable for the task through secondary screening.Finally use the greedy algorithm to determine the most suitable uninstall server based on the principle of minimum task response time,to obtain the final Uninstallation strategy and minimum response time.Finally,various performances of KMM and GMDC algorithms are evaluated through simulation experiments.The experimental results show that: Compared with some existing algorithms,KMM has a higher task unloading rate and lower task response time,which effectively improves to improve the user’s service quality.The GMDC algorithm mainly offloads the task to the vehicle,reducing the burden on the RSU and the cellular network at the cost of a small delay,and saving the cost of the operator.