Effective And Cooperative Resource Management For 5G/B5G Network

To deal with the challenges brought by new and complex network scenarios in 5G/Beyond 5G(B5G),we conduct the research on efficient and collaborative resource management for 5G/B5G networks in the thesis.Moreover,we mainly focus on proposing appropriate resource management model and corresponding algorithms to solve the issue of energy consumption and the requirement of ultra-low latency.As for the issue of energy consumption,we propose energy efficient resource management models based on single-antenna and multi-antenna heterogeneous networks(HetNets),correspondingly.As for the requirement of ultra-low latency,we introduce the mobile edge computing technique and mainly propose corresponding resource allocation methods to solve the network delay bottleneck with the limited energy consumption on the user/terminal side.Specifically,the contributions of the thesis are summarized as follows.(1)To solve the problem of energy consumption in the 5G/B5G network,we adopt the research idea of "throttling" for the single-antenna and multi-antenna HetNets and propose the corresponding energy-efficient resource management algorithms.As for the scenario of single-antenna HetNets,we consider the user experience and QoS requirements at the edge of the network and mainly focus on joint optimizing the spectrum allocation and power control to solve the extra energy consumption and the severe co-channel interference caused by the dense deployment of small cells.As for the scenario of massive MIMO enabled HetNets,we mainly focus on solving the energy consumption problems caused by massive MIMO deployment and consider the spectrum efficiency,traffic load distribution and other factors to propose the energy efficient joint optimization algorithm of power control and user association.(2)Moreover,we also use "open source" as the research idea,that is,we introduce the renewable energy to the power supply side of the base station to reduce the demand for grid power,and then study the corresponding resource management model.Considering the uncertain and intermittent output characteristics of renewable energy,we first describe traffic and energy models,and then formulate the user association problem as a multi-objective optimization problem to minimize the average waiting delay of users and on-grid power consumption.An iterative algorithm is proposed on both sides of the base stations and the users,which can dynamically adjust the traffic load distribution between the goal of maximize the renewable energy efficiency and minimizing network delay.(3)To satisfy the requirement of ultra-low latency in the 5G/B5G network,we introduce the edge computing to adjust the network architecture and mainly focus on reducing the network latency caused by the limited energy on the user/terminal side.Different from related works,the inter-user interferences caused by computation offloading demonstrate effective management in this paper.We formulate a joint computation offloading and resource allocation optimization problem to minimize the weight-sum delay of users under the constraint of inter-user interference and energy consumption.To improve the proposed method mentioned above,federated deep learning is introduced to propose the novel and distributed intelligent computation offloading algorithm.Specifically,we firstly use the Deep Neural Network(DNN)to learn and infer the mapping rule from the channel gains to the optimal computation offloading decision.And then we design distributed computation offloading algorithm based on federated learning(FL),which effectively reduces the algorithm execution time.