| In order to meet the access requirements of massive terminal devices,the deployment of 5G/B5G wireless networks tends to be intensive.This intensive trend of 5G/B5G networks has exacerbated the shortage of network resources.At the same time,the domestic and international community’s call for building a green communication network is getting higher and higher.On the other hand,as wireless technology evolves to 5G/B5G,its fundamental goal is to provide users with better and better quality of experience(QoE),as well as high and reliable quality of service(QoS)provision.The contradiction between the current shortage of network and energy resources and the ever-increasing QoE and QoS provision requirements are becoming more and more prominent.We must face up to this problem in the process of 5G/B5G network construction.In this thesis,the researches on the 5G/B5G wireless network green transmission technology,will be explored at two levels.The one is base station sleeping technology,and the other one is the wireless network resource management.The main work and contributions of this thesis are summarized as follows:(1)Research on maximizing sleeping capability based on QoS provisionConsidering the heterogeneous QoS requirements of Internet of Things(IoT)devices,such as the diverse coverage requirements and the diverse achievable data rate,the maximum sleeping capability of Information-Centric Internet of Things(IC-IoT)is explored.In the formulated nonlinear optimization problem,the sleeping probability acts as both an optimization variable and an optimization goal.Faced with this challenge,the properties related to sleeping probability and available bandwidth are analyzed.Then,a dichotomy-based nonlinear programming algorithm is proposed.The proposed algorithm divides the original optimization problem into two sub-problems.One of the sub-problems is related to bandwidth reallocation and traffic offloading strategies.The other one sub-problem is used to calculate the optimal sleeping probability.Last,the following facts can be found by analyzing the simulation results.The optimal sleeping probability is linearly proportional to the available bandwidth and inversely proportional to the user equipment density.Compared with the QoS guarantee requirements,the sleeping probability is more sensitive to the changes in channel status.(2)Research on distance-sensitive distributed repulsive sleeping strategyThe present centralized sleeping strategies generate huge communication overhead,have coverage holes,and reduce the network robustness.In order to overcome these problems,considering that the hard core point process constraints the distance for any two points,a distance-sensitive distributed sleeping strategy is posed.To characterize the performance of the proposed distributed sleeping strategy,utilizing the stochastic geometry theory,the analytical expressions of the base station sleeping probability,user coverage probability and average achievable data rate are derived.Finally,the coverage characteristics of three sleeping strategies are compared and analyzed,i.e.,the proposed distributed sleeping strategy,the existing random sleeping strategy and general repulsive sleeping strategy.Simulation results demonstrate that the proposed distributed repulsive sleeping strategy can provide more reliable coverage using the same energy consumption,compared with the random sleeping strategy and the general repulsive sleeping strategy.(3)Research on cross-layer resource allocation for multi-hop V2X communicationsUsing the V2X communication technology,the vehicles could play the roles of mobile relays in a multi-hop transmission communication system.Under such this system,users’ data queues are modeled.A long-term QoE optimization problem for users under the transmission system is studied.Faced with this long-term QoE optimization problem,an online algorithm for cross-layer resource allocation is introduced.The Lyapunov optimization is developed to transform the long-term optimization problem into a series of instantaneous sub-problems.These sub-problems involve the joint optimization for the rate control,power allocation and mobile relay selection under each time slot.On the one hand,the optimization problem of rate control is decoupled and independently optimized.On the other hand,a low-complexity price-based stable matching algorithm is proposed to solve the joint mobile relay selection and power allocation problem.Last but not the least,the simulation results verify that the queue backlog and network utility follow the trade-off relationship of[O(v),O(1/v)]under the proposed algorithm framework.(4)Research on the large-scale resource allocation for the Internet of Things network based on ADMMThe characteristics of decomposition and coordination for alternating direction method of multipliers(ADMM)could accelerate the speed of convergence.Considering this,three kinds of large-scale resources allocation problems in the IoT network are studied.The first one is a nonlinear fractional programming problem.For this problem,the Dinkelbach algorithm is introduced to convert such problems into equivalent convex large-scale resource allocation problems,then the classic ADMM with two variables is applied to solve the equivalent convex optimization problems.The second one is a large-scale resource allocation problem with high-dimensional variables and a large number of constraints.Then the Jacobian-ADMM algorithm is introduced to solve the optimization problem in parallel.The third one is a game-based resource allocation problem with high-dimensional variables and a large number of constraints.Faced with this,a two-layer iterative resource allocation algorithm is designed based on the Dinkelbach algorithm and the Jacobian-ADMM algorithm.The simulation results show that the three proposed algorithms have excellent convergence and adaptability to network scale expansion.Especially,the two-layer iterative resource allocation algorithm can linearly converge to the Stackelberg equilibrium point.(5)Research on two-timescale resource allocation for the 5G empowered industrial Internet of ThingsIn the communication system powered by hybrid energy,the energy harvesting(EH),electricity price,channel status and data arrival vary at different time granularities.Faced with this problem,a two-timescale resource allocation problem is formulated in the Industrial Internet of things(IIoT)network.Combining Lyapunov optimization,ADMM and matching theory,an online algorithm is proposed.This proposed online algorithm first applies the Lyapunov optimization to decouple the established long-term network performance optimization problem into a series of sub-problems,including the energy management problem at a large time scale,the rate control problem at a small time scale and the joint channel selection and power allocation at a small time scale.There are the sum constraints consisting of a large number of variables in the rate control problem.Thus,an ADMM-based low complexity rate control algorithm is proposed.Then,the joint channel selection and power allocation problem is converted into a one-to-many matching problem,which is solved by the proposed price-based quota-limited matching algorithm.The simulation results indicate that the proposed online algorithm can effectively reduce the peak-to-average ratio of data queue backlog and data arrival rate by 35.5%and 28.5%,and reduce energy costs by 48.23%.The proposed online algorithm can dynamically adjust resource allocation strategies and provide differentiated services for devices according to service priority and time-varying electricity price.Furthermore,the proposed algorithm reduces the computation time by 99%in exchange for 1%optimal performance degradation. |
PDF Full Text Request