Research On Joint Optimization Technology Of Spectrum And Computing Resources In Internet Of Vehicle

Internet of Vehicles(Io V)is one of the core technologies of future intelligent transportation,with a large amount of network access and computational offload demand.However,the scarcity of spectrum and computational resources in current wireless networks and the high complexity and fast time-varying characteristics of vehicular networks make the traditional resource scheduling methods have certain limitations.Therefore,in order to adapt to the multi-user and highly dynamic vehicular network environment,it is important to study an efficient and reliable network resource allocation scheme.This thesis focuses on joint spectrum and computing resource optimization techniques in vehicular networking,combines the application service requirements of vehicular network environment,proposes a model for spectrum and computing resource allocation in vehicular networking,and provides an optimal resource allocation strategy for vehicular networks.Firstly,the Multi-Agent Deep Reinforcement Learning(MADRL)based spectrum sharing algorithm for vehicular networks is proposed to address the problem that it is difficult for base stations to collect and manage transient channel state information in a highly dynamic vehicular network environment.The algorithm aims to maximize the network throughput under the constraints of vehicle communication delay and reliability,and uses the learning algorithm to improve the spectrum allocation and power selection strategies.The optimization problem is modeled as a Markov decision model,which enables vehicles to have the cognitive ability to select channels and power transmission autonomously to explore the optimal spectrum sharing strategy.Simulation results show that the average successful payload delivery rate of the proposed algorithm can reach 95.89%,which is 16.48% higher than that of the random baseline algorithm,and can quickly obtain the near-optimal solution,which has significant advantages in reducing the signaling overhead of the Telematics communication system.Secondly,a distributed joint task offloading and spectrum access algorithm based on MADRL is proposed for the problem of increasing mobile data traffic in Telematics that leads to computational task overload.The algorithm constructs the optimization problem as a multivariate coupled integer programming problem,introduces Unmanned Aerial Vehicle(UAV)technology and Multi-Access Edge Computing(MEC)technology,and designs an access policy and task offloading mechanism to minimize the system execution delay.Simulation results show that the proposed algorithm can maximize task offloading and reduce system execution delay with good convergence and scalability.At last,aiming at the problems of scarce spectrum and computing resources and inefficient resource allocation in Telematics,a rule-driven intelligent resource allocation scheme is proposed to achieve efficient and reliable allocation of spectrum and computing resources in Telematics through policy distribution,reasoning,and execution.In addition,this thesis designs a Rete inference algorithm based on sharing degree and B+ tree structure with the optimization goal of minimizing node redundancy and time delay,which improves the resource allocation efficiency of the system.Simulation results show that the proposed algorithm can achieve the accurate construction of the rule inference network through node sharing and B+ tree,which effectively reduces the network complexity and improves the rule inference speed compared with the traditional Rete algorithm.