Research Vehicular Ad Hoc Network Computing Unloaded

With the development of computer technology, network technology, there are more and more global mobile computing users and the speed of people’s life is becoming faster and faster. It makes the need for the mobile computing higher and higher. As the mobile devices have many features, such as limited resources, the poor battery, insufficient memory, etc, the tasks which require larger computing can’t be completed independently on a single node. Thus we need to share idle computing resources around, divide the task and then offload one part of the task to other resource nodes to execute. It not only makes full use of the fragmented resource blocks in the network, but also reduces the energy consumption and overhead of the vehicle node.The thesis studied the problem of computing offloading in Vehicle-to-Vehicle scene with high mobility in the Vehicle Ad Hoc Network and achieved the goal that the user can compute at anytime and anywhere. It introduced a computing offloading framework which supports the mobility of the mobile devices. It managed the resource information of vehicle node with high mobility, combined the partition tasks and the target node selection and realized the computing offloading in the Vehicle Ad Hoc Network. With a consideration of many factors, the thesis proposed a method of computing offloading target node selection. The method selects the vehicles based on the resource information and assigns the task size to the vehicle node according to the computing power of the node and the service time.It built a simulation environment with Eclipse by using Java language in the thesis. It also realized four target selection strategies and proved the feasibility and effectively of the multi-attribute target selection strategy through the experiment. By analyzing the simulation results, it can get that the multi-attribute target selection strategy has the least partition task numbers and lowest cost for the executions, the completion rate of the task is highest and the task will be completed earliest with the increasing number of the target nodes, but the completion time of the task will be longer than other target selection strategies.