Distributed Resource Management In Dense Wireless Networks

The increasing popularity of intelligent mobile devices and diversification of multimedia services have put tremendous pressure on 5G.To support the access of massive devices and explosive traffic demand,network densification has become a key technology in 5G,and dense wireless network is becoming the inevitable trend for wireless networks.In dense wireless networks,small base stations are densely deployed such that users are close to base stations.As a result,the spatial reuse gain is achieved and the network throughput can be improved.However,there are some obvious features in dense wireless networks,such as unplanned deployment of small base stations,severe interference among base stations,limited cache storage and backhaul link capacity,heterogeneous and uncertain traffic.As a result,traditional resource management methods cannot work well and fail to achieve the optimal match between resources and traffic.Hence,designing efficient resource management techniques such as user association and content caching to match resources with traffic optimally is crucial and worthy of investigation.However,in dense wireless networks,distributed resource management faces new challenges.For one thing,content caching in base stations greatly increases the dimension and complexity of resource management.Therefore,it is challenging to schedule communication and storage resources jointly,while satisfying resource constraints and user requirements.For another,the cooperation among base stations makes resource management much more complex.In addition,the resources of base stations are heterogeneous,contributing more difficulty in resource management in dense wireless networks.To address the new challenges in dense wireless network,considering the heterogeneity and spatial reuse of resources,we investigate the resource management in dense wireless networks,aiming at achieving the perfect match between network resource and user traffic and thus improving network performance(e.g.,load balancing,cache hit ratio,and delay).The contribution of this dissertation can be summarized as follows:· We propose a joint base station activation and user association scheme to match communication resources with user traffic and achieve load balance among base stations,while taming traffic uncertainty in dense wireless networks.We firstly exploit chance constraints to address the uncertain traffic arrival rate,and then transform them into deterministic constraints with the known mean and covariance,regardless of the dis-tributions.On this basis,a distributed base activation and user association scheme is proposed,where a cell-edge user can access multiple base stations opportunistically to avoid congestion.Through theoretical analysis,we show that the proposed scheme can converge fast and obtain the optimality gap between our proposed scheme and the optimal solution.Finally,simulation results validate the correctness of the theoretical analysis and demonstrate that our proposed scheme is near-optimal.We find that the proposed scheme is more robust to traffic uncertainty,and can alleviate the severe interference caused by network densification,reaping the spatial reuse gain to achieve load balance among base stations.· We propose a joint content caching and content recommendation scheme to maximize cache hit ratio,while meeting users' needs.By exploring cooperation among base stations,we can make full use of cache storage at base stations.Through optimizing content caching at base stations,we can match users' traffic with cached content.Similarly,via optimizing content recommendation,we can match cached content with users' traffic.Specifically,we are the first to propose a general content caching and recommendation framework.On this basis,we propose an online and distributed content caching and recommendation scheme with the aid of continuous time Markov chain.Through perturbation analysis,we investigate the estimation error on our proposed scheme and derive its low bound.Simulation results validate our theoretical analysis and the advantage of the proposed scheme.Through cooperative content caching and recommendation,we can overcome the limitations of backhaul links and cache storage,achieving the perfect match between network resource and user traffic.· We propose a joint content caching and user association scheme to achieve the match among communication resources(access and backhaul link capacity),storage resources,and user traffic,resulting in minimizing users' perceived delay.Considering both dynamic user population and natural tandem systems,we model content delivery in each base station as a tandem queue and thus derive the average download delay.Then,to minimize the delay,we optimize content caching and user association by exploring matching theory and dynamic programming,obtaining the near-optimal solution.Theoretical analysis demonstrates the convergence of the proposed scheme and effectiveness of the proposed scheme.Simulation results show that our proposed algorithm can significantly reduce the average download delay when compared with other schemes.As user association and content placement can both reduce the delay,we also demonstrate that content caching dominates user association when the backhaul is limited or user traffic becomes heterogeneous.Through optimizing content caching and user association jointly,we have more freedom to split user traffic into base stations to match with base station resources,such that the average download delay can be reduced further.