Research On Task Offloading Decisions For MEC-assisted Platooning

In the future network of C-V2 X,due to the limited computing resources of a vehicle terminal,it is impossible to solely rely on the vehicle itself to execute some computation-intensive applications and services.Simply executing tasks on a vehicle terminal will not only generate excessive computing load,but also consume too much energy,which increases the cost of the vehicular user and reduces the battery life of the vehicle.Mobile edge computing(MEC)is considered as one of the key technologies to strengthen the capability of vehicle computing,which has a large number of computing and storage resources,and users can reduce execution cost and energy consumption by offloading tasks to an MEC.C-V2 X can provide more safer driving on the road and more experiential services for the vehicular users.However,the diversity of task topology model,the randomness of task arrival and the high demand of resources for tasks execution result in more challenges to realize task offloading in the scenario of C-V2 X.Currently,the tasks offloading based on an MEC-assisted network framework has been studied,where it mainly concentrates on the cost of tasks offloading,tasks execution time,delay and other related aspects,but does not consider the optimal tasks offloading in a platooning.To deal with this problem,this thesis proposes a cooperative task offloading decision with minimum cost for an MEC-assisted platooning,combined with the model of the tasks.Each sub-task of the application is modelled as a linear topology and dependent on each other.Every member in the platooning and the MEC server have the opportunity to execute the tasks.This thesis presents optimization of the offloading cost based on the LARAC.The proposed approach can cooperate with an MEC server to execute tasks,and ensure that tasks are completed within a predefined time constraint and the cost of tasks execution is minimized,and is suitable for platooning scenario.Considering task offloading for the case of multiple vehicular users,a dynamic offloading algorithm based on Lyapunov optimization is proposed under the premise of ensuring the stability of the system in the thesis.According to the tasks initiated by vehicular users at each time,the queues model of tasks computing cycles are established.By employing Lyapunov optimization method,the task offloading strategy is presented,jointly considering energy consumption and execution time.Based on data size of the tasks,the offloading decision of all tasks is dynamically adjusted to determine whether they are executed locally,in a platoon member or in an MEC server.The proposed algorithm can effectively reduce energy consumption of tasks execution,maintain the queues stability of tasks computing cycles,and ensure that the tasks are completed within the deadline,which is one of the effective and feasible ways to realize dynamic offloading of tasks.