| Internet of Things(Io T)is able to greatly facilitate industry and people’s life by realizing connections between “people-to-things” and “things-to-things”.With the rapid development of wireless technologies,the fifth generation mobile communication networks would enable Io T applications,which provide more sustainable and stable energy,effective information transfer,flexible computing services.Recently,with the expansion of the scale of Io T and the increasement of Io T applications,Io T is also facing many challenges.Firstly,with the proliferation of Io T devices,it leads to huge energy consumptions.In addition,for low-power Io T devices that are not easy to change batteries frequently or are not suitable for battery deployment,how to deal with their energy supply problem is a major issue.Secondly,due to the lack of spectrum sub 6 GHz,new multi-user access and transmission technologies are also needed.Thirdly,in the face of diversified needs of Io T,to meet the personalized requirements of users is also a crucial problem.Especially,for the Io T devices with delay-sensitive computing tasks,it is a serious challenge to provide computing services that meets delay constraints in Io T systems.Thus,it is necessary to study Io T systems which can provide stable energy supply,improve spectral efficiency,and support users’ computing and communication requirements.With the basic goal of reducing energy consumption and increasing energy efficiency,and based on the optimization theory,this dissertation studies resource optimization of multi-user networks driven via advanced 5G/B5 G technologies.By means of system model construction,problem description and analysis,algorithm designs and complexity analysis,simulation verifications and comparisons,four innovative studies are carried out according to the classification of users with or without delay-sensitive computing tasks.The novelties and main contributions of the dissertation are including:1)Considering the non-linear features of EH and usage in the “harvest-then-use” protocol of RF EH,this work focuses on the system optimization designs with RF EH.In the SWIPT system,the system energy efficiency is maximized by jointly optimizing the transmit beamforming vector,artificial noise matrix and phase shift matrix.The efficient solution of the system is obtained by using a low-complexity algorithm based on semidefinite relaxation(SDR)and alternating optimization(AO).Simulation results show that the proposed scheme can effectively improve the system performance.In addition,in the WPCN system,by jointly optimizing the energy beamforming vector and time allocation,the system energy consumption can be minimized.Based on variable substitution and SDR technique,an effective method is proposed to obtain the efficient suboptimal solution.The optimality is verified by simulation results,and the effects of different schemes on the system performance are also compared.The system design using the nonlinear model can be closer to the practical nonlinear situation.2)Considering the non-linear features of storage in the “harvest-and-store-then-use”protocol of RF EH,this work studies the optimal designs in full-duplex(FD)and NOMA assisted SWIPT networks.By jointly optimizing the transmit beamforming and the power splitting(PS)factor,two optimization problems are formulated to minimize energy consumption with two kinds of channel state information(CSI)settings,respectively.Based on full CSI,SDR is used to relax the original problem.Then,a bilevel optimization programming(BLP)-based and a successive convex approximation(SCA)-based algorithms are proposed to obtain the conditional global optimal solution and the local optimal solution,respectively.After that,the closed-form solution under the single transmit antenna configuration is derived.Based on partial CSI,the probabilistic constraints are firstly transformed into deterministic constraints,and then the effective local optimal solution is obtained by applying SDR and SCA.Simulation results reveal that the proposed scheme can greatly improve the system performance.Besides,it is showed that the effects of nonlinear storage features on system performance are different from the results of the linear model.3)Considering the Io T devices with divisible computing tasks,this work discusses the optimal design of fog computing-assisted SWIPT systems.For the optimization design with fixed offloading time(FOT),the beamforming vectors and matrix,bandwidth allocation and offloading distribution are jointly optimized,while for the optimization design of optimized offloading time(OOT),not only the above resources are jointly optimized,the offloading time allocation is also considered.In the FOT design,the original problem becomes a convex problem after applying SDR,and then the semi-closed solution is obtained.In the OOT design,an algorithm based on penalty dual decomposition is adopted to obtain the local optimal solution.Simulation results reveal the tradeoff between accuracy and complexity under the two system designs.4)Considering the Io T devices with indivisible computing tasks,this works studies the fairness of multi-user in wireless powered cloud-fog systems by jointly optimizing the computing mode,as well as computation and communication resources.Due to the complexity of the formulated problem,this work firstly applies an algorithm based on generalized Benders decomposition to obtain the global optimal solution.For a given computing mode,the closed-form solution of the system can be achieved.Then,in order to strike a tradeoff between accuracy and computation complexity,this work also designs an algorithm based on penalized-SCA(P-SCA)to obtain an effective local optimal solution.The simulation results verify the tradeoff between accuracy and computation complexity. |
