Research On Joint Optimization Method Of Task Offloading And Resource Allocation Based On Priority-aware In The Internet Of Vehicles

In recent years,the number of motor vehicles is on a straight upward trend accompanied by economic development,and the data and computing tasks generated have also increased exponentially,making it challenging for vehicles to use their limited resources to perform intensive and sensitive tasks on the local CPU.Mobile edge computing(MEC)network appeared at the historic moment.It deployed servers on roadside basic units close to user vehicles to reduce time delay and energy consumption.In vehicle edge computing networks(VECN),the task transfer process from the user vehicle to the roadside basic unit is called task offloading.However,the resources of the MEC servers are not infinite,so how user vehicles choose to offload terminals and how edge servers allocate resources to minimize time delay and energy consumption are crucial to the research of mobile edge computing.Therefore,this paper divides the computing task into two states: integrity and cutting,and studies the joint optimization methods of task unloading and resource allocation based on priority under different states.The system environment of multi-user vehicle and multi-MEC server is constructed,and the task execution mode is analyzed,and the local computing model and computing unloading model are established,including network communication model,time computing model and energy loss model.This paper points out that tasks can be completely processed as unit "1" or tasks can be cut and processed separately.In the two states of task integrity and cutting,the above model is used to realize the model construction of time delay minimization and energy loss minimization,and the two optimization objectives are linearly weighted according to the priority of the task to form a fitness function,so as to facilitate the solution of subsequent algorithms.First of all,based on task integrity,a collaborative task offloading and resource allocation(CTOARA)joint optimization scheme is proposed,and an improved swarm intelligence optimization algorithm(IVA-GWO algorithm)is designed for the CTOMSA scheme to achieve the optimal solution.IVA-GWO algorithm improves the effective factor in the traditional grey wolf optimization algorithm,and proposes to carry out mutation operation according to the adaptive mutation factor to select the best offspring,which improves the development and exploration ability.Secondly,based on task cutting,an improved algorithm(CWL-PSO algorithm)combined with Levy flight.Specifically,chaos map is used to initialize the solution to make the candidate solutions uniformly distributed in the solution space.After the dynamic adjustment of inertia weight and learning factor is proposed to solve the problem that the particle search range remains unchanged,the velocity vector is updated in combination with Levy flight in order to ensure that algorithm has a minimum probability of falling into the local extremum and improve the accuracy of the algorithm.Finally,this paper designs multiple groups of tests under the two states of the task to check out the rationality of the algorithm.The experimental results show IVA-GWO algorithm is better than other swarm intelligence optimization algorithms when solving the problem of task offloading and resource allocation under the condition of task integrity,and the CTOARA scheme can better reduce time delay and energy consumption.When solving the problem of task offloading and resource allocation in the case of task cutting,the proposed CWL-PSO algorithm has better convergence and stability than other algorithms.