Resource Allocation And Offloading Decision For Vehicular Edge Computing Networks

With the rapid development of Internet of Things(Io T)and wireless communication technology,smart vehicles and vehicular new applications such as autonomous driving and assisted driving have been vigorously developed.These applications have the characteristics of being computationintensive and delay-sensitive.Vehicles with limited on-board computational resources are difficult to provide better quality of service(Qo S)for these applications.Mobile edge computing(MEC)has sufficient communication,computational,and storage resources,and is located at the edge of the network and close to users.MEC can not only solve the problem of increased delay and energy consumption caused by limited vehicular computational resources,but also solve the problem of excessive transmission delay and overhead caused by long-distance transmission of cloud computing.Therefore,offloading computation-intensive tasks to MEC servers is considered as a promising solution to improve vehicular tasks’ Qo S.However,MEC servers are still resource-constrained,offloading to MEC servers is sometimes unable to meet the Qo S of vehicular computation-intensive tasks,especially for those under a high traffic flow.If there is a large gap in the number of association users between different MEC servers at the same time,it will cause unbalanced load and low resource utilization.Misbehaving vehicles result in lower network efficiency and even disrupt network in vehicular edge networks(VEC).Therefore,how to design a reasonable user association and resource allocation scheme,use idle resources in the network to expand MEC servers’ resources to help the computation-intensive vehicles to improve their Qo S and recognize misbehaving vehicles in VEC are the focus of this paper.Therefore,this paper consists of the following parts:(1)Joint Offloading Decision and Resource Allocation for Vehicular Parked-Edge Computing Networks: A Contract-Stackelberg ApproachDue to the limited resources of the MEC server and a large amount of free resources in the parking lot(PL)that have not been utilized.We introduce the Vehicular Parked-Edge Computing(VPEC)paradigm to formulate a three-stage contract-stackelberg offloading incentive mechanism in order to tackle this problem.The paradigm has three levels: a PL level,an operator of MEC level and a vehicle level.Due to information asymmetry between PL and parked vehicles(PVs),We use contract-based incentive mechanism to motivate PVs to contribute their idle computational resources.The PVs are classified into different types according to their parking time,and PL offers different contracts to different types of PVs.Then,based on Stackelberg game,we develop a pricing model that maximizes the utility of three players,including vehicles,operator,and PL.We use backward induction method to solve this three-stage problem,and give the closed-form expressions of the optimal strategies for each stage.Simulation results demonstrate the feasibility of the proposed incentive mechanism and reveal the changing trend of optimal strategies in each stage when traffic density changes.We also reveal the trend of energy consumption and the offloading decision of the computation-intensive vehicles with various parameters in simulation results.(2)Reputation-based Joint User association and Resource Allocation for Vehicular Edge Computing Partial offloading NetworksBased on the previous work,we further consider the user association problem.For MEC server’s limited resources,possible unbalanced load,and the potential safety issues,we establish vehicular reputation model to recognition misbehaving vehicles and propose a reasonable user association,resource allocation,and partial offloading scheme based on vehicular reputation to solve the above problems.We propose to jointly optimize user association,local and MEC servers’ computational resource allocation,uplink transmit power control,and offloading ratio to minimize the weighted sum of the system’s delay and energy consumption for computing.This optimization problem is a mixed-integer and non-linear programming problem,and hard to obtain a global optimal solution.We decompose the optimization problem into user association subproblem,joint resource allocation subproblem,and offloading ratio subproblem.The user association subproblem is solved by a many-to-one matching game approach based on vehicular reputation.The joint computational resource allocation subproblem is solved by Lagrangian multiplier method.The offloading ratio subproblem is a linear problem.The three subproblems iterate each other until convergence,and a suboptimal solution is obtained.Simulation results show that the algorithm proposed in this paper has better performance compared with the two algorithms that are only local computing and only MEC computing.