Optimization Of Task Offloading Algorithm Based On Mobile Edge Computing In The Internet Of Vehicles

In recent years,the intelligent transportation industry has developed vigorously,and the processing performance of tasks requested by vehicle terminals in the Internet of Vehicles has been continuously improved.However,the traditional cloud computing big data system architecture cannot meet the low latency requirements of Intelligent Traffic System(ITS)applications.For its drawbacks,integrating Mobile Edge Computing(MEC)into the communication network architecture can improve the processing capacity of complex tasks in the Internet of Vehicles and reduce the processing delay of tasks.However,most of the conventional task uninstallation methods are executed on the MEC server,and there is no clear classification of the types of uninstall tasks.In this case,how to design an effective offloading decision to make full use of local resources and server resources has become an urgent problem to be solved.This paper studies the optimization of computing offloading algorithms based on Mobile Edge Computing in the Internet of Vehicles environment.The main research contents are as follows:First of all,in view of the current problem of insufficient computing resources for mobile vehicle terminals to meet the requirements for reliability and low latency.On the one hand,Software Defined Network is used to achieve centralized control and resource allocation.On the other hand,computing resource allocation and task offloading decision variables can be used as adjustable parameters,then it is modeled as an optimization problem of minimizing the weighted sum of network delay and energy consumption.Since this problem is an NP-hard optimization problem,the sub-optimization scheme is used to decompose it into two sub-optimization problems for solution,and a task offloading algorithm based on branch and bound method is proposed.The algorithm firstly uses the Lagrangian multiplier method to solve the computing resource allocation problem,and secondly,it prioritizes the maximum tolerable delay of the safety message task and uses the branch and bound method to obtain the optimal offloading decision.The simulation results show that the offloading algorithm based on the branch and bound method proposed in this paper has reduced processing delay compared with the traditional task offloading algorithm and the energy-optimized offloading algorithm.Secondly,considering the shortcomings of lack of flexibility caused by modeling task offloading as an alternative problem of local computing or offloading to the MEC server in existing works,this thesis assumes that the tasks can be split and then optimizes task splitting factors and offloading decisions at the same time.The MEC system goal is modeled as an optimization problem that minimizes the weighted sum of time delay and energy consumption.Next,a partial offloading algorithm based on Q learning is proposed.Specifically,the algorithm first classifies vehicle terminals according to delay and computing resource requirements,and then offloads the extremely timesensitive tasks on the local vehicles and the tasks with high demand for computing resources to the MEC server for processing.The remaining tasks are partially offloaded,and Q learning is used to get the best offloading decisions.The simulation results show that the partial offloading algorithm based on classification proposed in this thesis greatly reduces the complexity of the algorithm.