Research On Mobile Data Offloading Methods For Vehicle Edge Networks

With the development of information technology and automotive industry,the demands for mobile data from vehicle users are growing day by day.Due to hardware resource constraints,a large number of mobile data demands from vehicle users increase the load on the cellular network base stations greatly,which reduce the Quality of Experience(Qo E)of vehicle users.In order to relieve the load on the cellular base stations as well as guarantee the Qo E of vehicle users,researchers propose the data offloading technology through Vehicular Ad Hoc Networks(VANETs).However,the current works lack of the research of the impacts caused by the vehicle mobility(such as the relative speeds and directions between vehicles).Meanwhile,most of the studies do not give consideration to both of the economic benefits of mobile network operators(MNOs)and the Qo E of vehicle users.Therefore,from the perspectives of the economic benefits of MNOs and the Qo E of vehicle users,impacts of the vehicle mobility are taken into consideration in this thesis.Based on this,the data offloading methods through VANETs under two different traffic conditions are studied in the thesis.(1)In the case that the speeds of vehicles are stable and the relative speeds between vehicles are small,a V2 V offline data offloading method based on vehicle mobility is proposed in the thesis.The impacts of the vehicle mobility are considered in the thesis.The data are transmitted between vehicles in a multicast manner for the sake of fairness.Meanwhile,from the perspectives of MNOs and vehicle users,the benefits of MNOs as well as the Qo E of vehicle users are taken into consideration in the thesis.The problem is formalized as the minimizing cost of the MNOs problem and the maximizing overall data utility of the data requesting vehicles problem.In the process of solving the problem,the minimizing cost of the MNOs problem is transformed to the multiple backpack problem,then a algorithm based on dynamic programming and backtracking is proposed to solve the problem.By binary splitting,the performance of the algorithm is improved.The maximizing overall data utility of the requesting vehicles problem is a mixed integer nonlinear programming problem.So the problem is converted to a mixed integer linear programming problem,and solved by the branch and bound method to obtain a mobile data offloading strategy.(2)In the case that the speeds of vehicles are unstable or the relative speeds between vehicles are large,a V2 V online data offloading method based on vehicle mobility is proposed in the thesis.The network service process is divided into continuous and equal-sized time slots.Referencing to the Manhattan model,a vehicle movement model is built in the thesis.From the perspective of vehicle users,a data utility model and a data requesting vehicles cost model are established.The problem is formalized as the maximizing overall satisfaction of the data requesting vehicles problem.Finally,in each time slot,a genetic algorithm is used to solve the maximizing overall satisfaction of the data requesting vehicles problem to obtain a mobile data offloading strategy.And then,the performance of the algorithm is enhanced by improving the algorithm.(3)Experimental simulations are performed on the V2 V offline data offloading method based on vehicle mobility and the V2 V online data offloading method based on vehicle mobility.The experiment results show that vehicle mobility has a great effect on mobile data offloading,and the mobile data offloading methods proposed in the thesis are effective.