Research On Task Offloading In Vehicular Edge Computing

In vehicular edge computing system,in order to support the execution of emerging computation-intensive and delay-sensitive applications,each vehicle can offload its computation tasks to RSU or cloud for execution.The process of migrating these computing tasks to obtain more resources is defined as task offloading.In this thesis,we study the problem of computation offloading.When offloading the computation tasks to RSU,the computing resources in edge nodes are fully utilized to optimize the execution delay of computing tasks and the execution cost of the edge nodes.The main research points of this paper and the research results obtained are as follows:(1)Aiming at the constraint of limited computation resources of RSU,we focus on the computation resource allocation strategy of RSU and the offloading decision of vehicles.By studying the interaction between RSU that allocate computation resources and vehicles that offload their applications,we model the problem as a multiple-leader and common-follower Stackelberg game problem,and prove the existence of nash equilibrium of the Stackelberg game.We also propose the local and edge equilibrium computing algorithm to compute a nash equilibrium,and propose the computation resource allocation strategy of RSU and the offloading decision of vehicles.(2)Aiming at the unbalanced tasks of RSU to execution,we focus on the cooperative offloading decision between RSUs.We consider cooperative computing between RSUs to make full use of the computing resources on each RSU,and optimize the delay cost and the energy cost of RSU.We model the cooperative computing problem in the vehicular edge computing as a non-cooperative game,and prove the existence of nash equilibrium of the game.We also propose the cooperative equilibrium computing algorithm to compute a nash equilibrium,and propose the cooperative offloading decision between RSUs.In the research of this paper,numerical simulation results are also performed on the above two issues.The comparison result proves the algorithm proposed in this paper outperform the benchmark approaches.