High-efficiency Energy-saving Resource Management Techniques For Wireless Cellular Networks

The explosive increase in mobile devices and ubiquitous wireless services has incurred significant energy consumption.The 5th generation mobile communication system(5G)has explicitly pointed out that it will adopt various techniques to comprehensively improve the energy efficiency of the whole wireless communication system.From the perspective of network,exploiting new energy sources(e.g.,wind and solar energy)can solve the high energy consumption problem of wireless networks fundamentally.However,different from the traditional grid power,the renewable energy often has the property of random fluctuation.Thus,inappropriate energy and traffic allocation strategy will degrade the energy conservation gain of the renewable energy and result in the loss of the network performance in severe case.Therefore,under the constraints imposed by the dynamic variation of the renewable energy and wireless traffic in both space and time,designing appropriate resource management techniques is critical for fully exploiting the renewable energy and reducing the power consumption of the network.On the other hand,to support massive connections,5G will employ multiple spectrum sharing techniques(e.g.,Non-Orthogonal Multiple Access and Device-to-Device),the combination of which leads to more serious interference among mobile devices.How to coordinate the interference among devices and thus reduce their power consumption via appropriate spectrum assignment and power control is a serious challenge for the design of the high-efficiency energy-saving resource management algorithms.Motivated by the above reasons,this dissertation devotes to investigate high-efficiency energy-saving resource management techniques for cellular networks.The research is carried out from two aspects,i.e.,network side and mobile device side energy-saving techniques.On the network side,we design energy and traffic dynamic scheduling algorithm and coordination algorithm for the energy-harvesting based wireless cellular networks,thereby reducing the grid power consumption caused by the base stations.On the mobile device side,we design a unified spectrum and power allocation algorithm for the wireless cellular networks with NOMA to coordinate the interference among devices and reduce their transmit power.The main achievements of this dissertation are summarized as follows:1.We proposed an energy and traffic dynamic scheduling algorithm to minimize the long-term grid power consumption in the single-cell downlink cellular network with hybrid energy supplies.This problem was formulated as a stochastic network optimization problem,consisting of the resource allocation in MAC(Media Access Control)layer and queue management in network layer.Based on the Lyapunov optimization technique and traditional optimization theory,we designed an effective energy and traffic dynamic scheduling algorithm.Different from the traditonal online algorithms,our designed algorithm can adjust the control policies(admission control,subcarrier assignment,power allocation,and transmit time optimization)according to the current system state but without requiring the statistical information of any random process(i.e.,energy harvesting,traffic arrival,and channel state).In addition,we proved via theoretical analysis that the proposed algorithm can converge to the global optimum asymptotically at the cost of queue backlog increase.Through adjusting parameter,our algorithm can strike different tradeoff between grid power consumption and queue backlog.Finally,simulation results verified the correctness of the theoretical analysis and the effectiveness of the proposed algorithm.Our proposed algorithm has the advantages of low complexity and easy implement that is beneficial for its application in actual system.2.We proposed an energy and traffic coordination algorithm,with the objective to minimize the total grid power consumption in the multi-cell downlink cellular network with hybrid energy supplies.In particular,we combined the smart grid with the cellular network to coordinate the spatial distribution of the renewable energy and mobile traffic.Firstly,for centralized networks,we devised a centralized energy and traffic coordination algorithm.Specifically,by jointly optimizing the spatial distribution of the renewable energy and mobile traffic,the centralized algorithm achieves a good match between the energy supply and demand at each base station,such that the grid energy expenditure of the whole system is greatly reduced.For distributed networks,we further proposed a three-phase distributed control policy: load-aware user association,intra-cell power and resource allocation,and inter-cell energy cooperation.The distributed control policy can be independently executed by the base station and users only with their local information.Finally,simulation results indicated that our algorithms have the advantages of good convergence,effectiveness,and flexibility.Comparing with the existing algorithms,our algorithm can achieve a better match between the energy supply and traffic distribution,thus fully exploiting the spatial diversity of the renewable energy to reduce the grid power consumption.Our research is a unified study on the smart grid,energy harvesting,and cellular network,which can provide theoretical guideline for the design of green communication system.3.We designed a unified spectrum and power allocation algorithm for mobile devices,aiming to reduce the power consumption of mobile devices in uplink D2 D and cellular hybrid network.To support the massive connections,5G will adopt multiple spectrum sharing techniques,e.g.,SCMA(Sparse Code Multiple Access)and D2 D.In D2 D and cellular hybrid network,D2 D users employ OFDMA(Orthogonal Frequency Division Multiple Access)and cellular users employ SCMA,which can greatly improve the network capacity.However,the complicated spectrum sharing pattern also leads to serious interference,which further boosts up the power consumption of mobile devices and degrades users' QoE.To tackle this problem,we jointly considered the mode selection,resource allocation,and power control to design a unified resource management scheme for D2 D users and cellular users.Firstly,we analytically got the optimal resource-sharing mode(i.e.,dedicated mode or reuse mode)for cellular users and D2 D users based on the mapping rule between SCMA codebooks and OFDMA resource blocks.For the dedicated mode,we reformulated the resource allocation problem as the flow issue in graph theory and designed the optimal algorithm based on graph theory and optimization theory.For the reused mode,a sequential resource allocation algorithm was put forward based on an interference aware graph.Then,an iterative power control algorithm was adopted to further optimize the transmit power for all users.Finally,simulation studies indicated that our algorithms can well coordinate the cross-layer interference between cellular users and D2 D users,thereby greatly reducing the transmit power of mobile devices.