Joint optimization of recommendation and caching based on user preference prediction

The development of the Internet things brings exponential growth wireless traffic, which puts great pressure on backhaul link. proactive caching some contents in edge device mobile network can effectively reduce repeated transmission same and relieve burden Moreover, introduction recommendation mechanisms reshape user's request improve cache hit ratio. However, optimization decisions is highly dependent users’ preference information for files. Here, a joint algorithm based prediction with multiple base stations cooperative proposed. To efficiency, Deep Crossing model adopted to predict preferences. Under constraints capacity, quantity bandwidth, an problem minimize total delay system formulated. Then, NP-hardness proposed proved it decoupled into three sub-problems, namely recommendation, user access sub-problems. Simulation results show that system.