Random Caching Of Ultra-dense Networks With Multi-antenna Transmission And Base Station Cooperation

To satisfy the exponentially growing requirements of the mobile Internet services,future communication networks need to exploit the data characteristics from many aspects to improve the network performance.Recently,caching strategy is proposed as a promising method to capture the demand characteristics of the vedio contents.More specifically,by storing the popular files in the cache of the base stations(BSs)during the off-peak time,caching strategy can greatly reduce the load and the delay of backhaul during the on-peak time.As one of the caching strategies,random caching strategy stores different file sets in the caches of different BSs,and thus,the number of files stored in the edge cache from all BSs is larger than the cache size of one BS.In this way,the random caching strategy guarantees that as many as possible users can obtain the required files from the edge cache.However,in the random caching strategy,the users are associated with the closest small base stations(SBSs)storing the requested file,but not the geometrically closest SBSs,and thus,the transmission quality may be deteriorated.Multi-antenna transmission and base station cooperation can enhance the received signal and manage interference by using the spatial freedom of degrees from single or multiple base stations,which is an effective method to improve the file transmission quality of random caching strategy.Motivated by these reasons,we investigate random caching-aided ultra-dense networks with multiantenna transmission and base station cooperation.The main resuts and contributions of this paper are as follows:Firstly,we propose a joint backhaul transmission and random caching strategy for the backhaul-limited multi-antenna small cell networks.To enhance the weak received power in random caching strategy,we utilize the maximum ratio transmission beamforming method to provide maximal array gain.Based on this beamforming method,we use the cache and capacity-limited backhaul at SBSs to provide users with required contents.Assume that the distribution of the nodes in small cell networks follows Poisson point process(PPP)and the channel coefficient follows complex Gaussian distribution,we derive the closed-form exact expression,an upper bound of the success probability in the general scenario as well as a simple asymptotic expression in high user density region.Based on the analytical results,we then consider the success probability transmission maximization by optimizing the design parameters in the cache-enabled multi-antenna small cell networks.After analyzing the optimal properties,we obtain a local optimal solution with low complexity by reducing the search region of the cached file set.To further simplify the optimization,we solve an asymptotic optimization problem in the high user density region,using the simple asymptotic expression as the objective function.Numerical simulations show that our proposed joint backhaul transmission and random caching strategy outperforms existing caching schemes.Moveover,with the increase of the antenna number,the optimal caching distribution is more dispersive,which guarantees that the users are more likely to obtain the required files from the edge cache.Secondly,we propose a tier-level random caching strategy spatial division multiple access(SDMA)multi-antenna HetNets.To enhance the network throughput in random caching strategy,we utilize the zero forcing beamforming method to provide multiplexing gain.Based on this beamforming method,we use the cache from the BSs in different tiers to provide users with required contents.Assume that the distribution of the nodes in small cell networks follows homogeneous independent Poisson point process(HIPPP)and the channel coefficient follows complex Gaussian distribution,we derive the closed-form exact expression and upper bound of the network throughput in the general scenario.We then consider the high target signal-tointerference ratio(SIR)scenario and obtain a simple asymptotic expression of the network throughput.Based on the analytical results,we consider the network throughput maximization problem via optimizing the tierlevel caching distribution.For a general scenario,we obtain a local optimal solution and a simple approximate asymptotic optimal solution in the high SIR threshold region.For a special case where each BS serves the largest number of users per resource block,we numerically solve a non-convex problem and propose a psuboptimal caching strategy by approximation,where a water-filling-structured closed-form result is obtained in each iteration.Numerical simulations show that our proposed tier-level random caching strategy outperforms existing caching schemes.Thirdly,we propose a joint interference nulling and random caching strategy for the multi-antenna small cell networks.To mitigate the strong interference in random caching strategy,we utilize the coordinate zero forcing beamforming method to null the inter-cell interference.Based on this beamforming method,we use the cache at SBSs to provide users with required contents.Using tools from stochastic geometry,we characterize the tradeoff between the random caching and the interference nulling by showing that the interference nulling cost is linear with the content diversity gain.We derive a closed-form expression of the success probability,as well as a simpler upper bound.Based on the analytical results,we jointly optimize the caching distribution and the interference nulling coefficient via the block coordinate ascent(BCA)method.Numerical results reveal that our proposed IN-aided random caching(INRC)strategy outperforms previous caching strategies,including a random caching strategy with an other IN scheme.Moreover,with the increase of the antenna number,the optimal interference nulling coefficient increases,which guarantees that the SBSs would null more interference for the users in other cells.Finally,we propose a joint coordinate transmission and random caching strategy for the backhaul-limited multi-antenna small cell networks.To enhance the weak received power and mitigate the strong inter-cell interference in random caching strategy,we utilize the coordinate transmission to provide macro diversity gain.Based on this coordination method,we use the cache at the SBS and the backhaul at the MBS to provide users with required contents.We model the locations of the MBSs and the helpers as a Poisson point process(PPP)and a Poisson hole process(PHP)to capture the spatial correlation between different tiers of BSs.We adopt a fitted Poisson cluster process(PCP)to approximate the PHP by matching the firstorder and second-order statistics and then derive explicit expressions of the average spectrum efficiency.Building upon the analytical results,we then design the caching resource allocation to maximize the average spectrum efficiency.Consider the scenario when the required file number is larger and the requirements are concentrated,we utilize a reasonable continuous function to approximate the file request probability,and then transform the mixed-integer caching resource allocation problem into a continuous optimization problem.After analyzing the Karush—Kuhn—Tucker(KKT)conditions,we then derive a closed-form solution for the approximate optimization problem.Numerical results reveal that our proposed joint coordinate transmission and random caching strategy outperforms previous caching strategies.