Research On Edge Cache Strategy Of Internet Of Vehicles Based On Reinforcement Learning

In recent years,with the development of the Internet of Vehicles and wireless technologies,and the continuous emergence of in-vehicle infotainment applications,the demand for content in the Internet of Vehicles is also increasing at an alarming rate.In order to meet such a high content demand in the Internet of Vehicles,edge caching technology has received extensive attention.By deploying cache resources on the edge of the network,such as base stations and roadside units,and caching some popular content on edge nodes to provide content download services,the pressure on the core network can be effectively relieved and content request delays can be reduced.However,due to the limited cache resources of edge nodes,uneven distribution of content requests,and dynamic changes in network topology caused by high-speed vehicle movement,it is necessary to develop effective cache content placement and delivery strategies for edge nodes.Therefore,this thesis considers the impact of road traffic conditions and user experience,and studies the collaborative caching strategy of edge nodes.First,this thesis considers the cooperative caching strategy of the macro base station and the roadside unit in the V2 I communication scenario.In order to study the impact of road traffic conditions on the caching strategy and how to allocate resources to provide content services for moving vehicles,this thesis establishes a dual time scale model.The large time scale mainly considers the different road traffic conditions in each area.Through content placement,some content requests can be offloaded to neighboring areas,achieving load balancing,and maximizing the utility function composed of download revenue and storage cost.The small time scale mainly considers the current request status of the content request vehicle,that is,the remaining acquisition content size and the remaining acquisition time.By specifying connection and bandwidth allocation decisions,the utility function composed of download revenue and transmission cost is maximized.Finally,the deep deterministic strategy gradient algorithm is used to solve the problem.The simulation results show that the proposed edge cooperative caching strategy has better performance in terms of system utility,offloading traffic and successful download rate.Secondly,this thesis considers the cooperative caching strategy of roadside units and mobile vehicles in the V2 X communication scenario.Using the cache in the vehicle can not only increase the capacity of the cache system,but also expand the service range of the cache and further improve the performance of the cache.For the in-vehicle cache,this thesis adopts the storage relay scheme,so that the relay method enables the cache content in the vehicle to remain in the considered area.In order to study the impact of user experience on caching strategies,this thesis combines the maximum allowable download delay of content to establish an index to measure user experience.The maximum acquisition delay of different content corresponds to different user experiences,and the greater the maximum acquisition delay of content,the greater the probability that the user can successfully download the completed content from the edge.Therefore,this thesis designs a cooperative caching strategy based on roadside units and vehicle caching to maximize the edge successful download rate.The simulation results show that the proposed strategy can effectively improve the successful download rate at the edge.