Research On Active Caching Scheme Based On Edge Computing In Internet Of Vehicles

With the explosive growth of Internet of Vehicles(Io V)data and the low latency requirements of vehicles on service requests,edge caching technology has become one of the hot spots in Io V research.At present,most existing research of the Io V caching only discussed the impact of the number of file requests on the caching scheme,and ignored the effects of the Io V environment,which leads to the fact that the more important files for vehicle driving are usually not cached in the limited storage space.At the same time,most research of the Io V caching only analyzed the situation where each vehicle only requests one file at each time,and ignord the situation where each vehicle can request multiple files at the same time,which leads to the poor practical applicability of the cache model.This thesis mainly focuses on the above two issues,studying the active caching scheme based on edge computing in the Io V.The main contents are follows:Firstly,considering that there are some files in Io V which have a greater impact on vehicle driving,however,are requested a few times,a caching scheme based on file values is proposed.In this caching scheme,the number of file requests and file priorities are cosidered.Besides,taking into account the system replacement cost and aiming to maximize file values,the proposed caching scheme is realized through the deep Q learning algorithm.Finally,the simulation results show that compared with the caching strategy which only considers the number of file requests or LRU caching strategy,the proposed caching scheme has a higher percentage of hit values and higher long-term caching efficiency,the more important files to vehicle driving can be cached in a limited cache space as many as possible.Secondly,considering that mobile vehicles may request multiple files at the same time,this thesis proposes a file set request caching scheme based on Markov transition probability,the optimization goal of which is to minimize the average request delay of vehicles.Then,this thesis constructs the sub-modular function of the optimization objective,and analyzes the system performance of the greedy algorithm.Finally,the simulation results show that as the number of files in the requested file set increases,the proposed caching scheme has a lower average request latency which is better than random caching algorithm.In addition,when the vehicle transfering probability changes,enlarging the vehicle requested file set can make the average delay time of vehicle requests in the entire macro base station cache area more stable.