Research On Caching And Resource Management In Dense Wireless Networks

Dense wireless networks could significantly improve network capacity by increasing the spatial reuse of spectrum resources.In dense wireless networks,user service experience would be improved due to the application of edge caching further shorten the distance between users and the required content.However,the application of edge caching changes the statistical characteristics of interference,which causes the interference not only depend on the location of base station(BS)but also on the location of content placement.As a consequence,the interference become more complex,difficult to control and even deteriorates the network capacity.Therefore,in dense wireless network,how to accurately describe the impact of edge caching on the statistical characteristics of interference,and how to reasonably guide and suppress interference is the key to improve the network performance.For the above reasons,we use network spatial throughput(ST)to characterize network capacity.In order to effectively improve dense wireless network information carrying capacity,we analyze the performance of the caching-enabled dense small cell network(CSCN)by using the theory of random geometry and optimization theory.Further,we maximize ST by the joint design of content caching and retrieving.The contribution of this paper can be summarized as follows:1.In order to improve ST,we joint design the optimal caching and retrieving(OCR) strategy,which achieves the optimal adaptation of storage resources and backhaul resources and ensures ST continue to grow steadily with the increasing of small cell base station(SBS).Specifically,a critical SBS density is derived from the optimization result,beyond which optimal content retrieving probability q_r~*and optimal cache hit probability q_h~*satisfy equation (?).And we could control backhaul data traffic by the above equation.The simulation results indicate that network ST stays stable growth by the joint optimization with the increasing of SBS density.2.In this paper,we propose a power control algorithm based on deep Q learning.The interference is effectively managed by predicting the interference of adjacent time slots,which makes ST is increased by 50%and guarantees the continuity of user service quality.Specifically,we investigate power control mechanism of SBS,which is defined as a random game process.As a learning agent,each SBS could accurately predict interference by capturing the correlation characteristics of interference in adjacent time slots.Further,each SBS manage interference via power allocation.The simulation results show that our proposed algorithm could increase ST up to 1.5 times when SBS deployment density is large.