Research On Edge Caching Strategy In The Internet Of Vehicles Based On Minimizing Energy Loss And Delay

Due to the rapid growth of the number of vehicles and large amount of data generated by various vehicle services,edge caching has gained wide attention.Vehicles and roadside infrastructure as network edge nodes can provide other nodes with data storage resources,computing resources and low-latency wireless connections,which can avoid high latency and network congestion caused by caching content from the core network.However,considering the high-speed movement characteristics of vehicles and the diversity of cache deployment,it is important to formulate efficient cache content placement and distribution strategies.Taking this as a starting point,this paper studies the placement and distribution of caching content in V2 V and V2 X communication scenarios,whose goal is minimizing cache delay and maximizing energy efficiency.First,in the V2 V scenario,based on the queuing theory and the actual traffic scenario,the probability of establishing a connection between the vehicle and other vehicles or base station is derived using a two-dimensional Markov process.In order to minimize the energy consumed by entire system,including transmission energy and cache energy,we have developed a cache scheme.The simulation results show that the proposed scheme has lower energy consumption efficiency than offline cache strategy.To reduce delay,the edge caching architecture of D2D-enabled V2 V communication is established.According to the size of distance,vehicles are divided into B-UEs and V-UEs.V-UEs need to reuse the spectrum of B-UEs for content transmission.In the cache content placement phase,the type of vehicle caching will be determined;in the cache content distribution phase,the method of reusing B-UEs spectrum will be determined to minimize the delay of entire system during caching.In the simulation process,the Hungarian algorithm and the continuous convex approximation algorithm are used to solve the integer non-convex optimization problem.The simulation results show that compared with the random cache strategy,the delay can be effectively reduced.Secondly,the V2 X caching architecture was established where the base station,RSUs and vehicles cooperate to cache,in order to achieve the goal of minimizing energy consumption,we proposes an algorithm to cache partial files.Compared with caching entire popular files and random caching files,the energy consumption of our algorithm is effectively reduced.No matter which edge node the vehicle establishes a connection with,only a certain percentage of content is cached in the current time slot.In order to solve this long-term mixed integer linear programming problem and to efficiently process a large amount of data information,this paper proposes a content caching strategy based on the deep deterministic policy gradient algorithm,by placing in advance and using the LRU algorithm to dynamically update the content of RSU s and vehicle storage,aiming to reduce delay,and the deterministic strategy of DDPG and the characteristics of small batch gradient descent also improve the convergence speed of the proposed algorithm.