Research On Joint Task Offloading And Resource Allocation Algorithm In Vehicular Networks With MEC

In recent years,the number of motor vehicles is growing rapidly,which has brought about the problems such as high rate of accidents and traffic congestion.In addition,people’s demand for diversified mobile infotainment services is also increasing.Empowering the Internet of Vehicles through emerging digital technologies such as 5G,big data,and artificial intelligence can provide smarter,safer,and more reliable driving services.However,the computation-intensive vehicular applications put forward higher requirements for the computation capability of the network,especially for the delay-critical autonomous driving services,ultra-low latency and ultra-high reliability must be achieved.The remote transmission of cloud computing and the limited resources of vehicles make it difficult to meet the computation requirements of such applications.Mobile edge computing,which deploys computation resource in the access network,is regarded as an effective solution to this problem.Edge computing has become the gordian technique to building an intelligent perception access network.Correspondingly,offloading decision and resource allocation have been the focus of relevant research.In this paper,the computation offloading problem under two typical network architectures in the vehicular edge computing network is researched,and the offloading decision and resource allocation are jointly optimized to achieve better network performance.In the single edge server vehicular networks,for streaming data processing applications,a partial offloading model which takes into account both delay and energy consumption is established.In the networks with limited communication and computation resource,the problem of minimizing system energy consumption under the maximum tolerable delay constraint is studied by considering jointly the offloading proportion decision and the resource allocation,so as to help build a green edge network guaranteeing the quality of service.Then,based on the intelligent optimization idea,a joint offloading and resource allocation strategy based on improved hybrid particle swarm and simulated annealing(IHPS-JORA)is proposed.Finally,the simulations verify that IHPS-JORA has better convergence and can achieve lower system energy consumption while taking up less computation resource of edge server.In the multi-edge server vehicular networks deployed in hotspots,for delay-critical security applications,in order to meet the real-time requirement,a delay model for full offloading is established.The communication resource pool sharing scheme is introduced to improve spectrum utilization.The problem of minimizing the average execution delay by optimizing jointly the integer offloading decision and resource allocation in the system with limited communication and computation resource.Then,based on the idea of bi-level programming,a joint offloading and resource allocation strategy based on improved bi-level particle swarm programming(IBLPP-JORA)is proposed to solve the non-convex problem.Finally,the simulations verify that IBLPP-JORA can effectively reduce average execution delay and improve the quality of experience of edge network.