On The Study Of Radio Resource Optimization For Green Heterogeneous Cellular Networks

After about a decade of research on the area of green wireless communications driven by economical and environmental considerations,reducing on-grid brown energy consumption and improving energy efficiency(EE)have become key performance metrics for designing current and future wireless communication systems.Actually,EE has become one of the key pillars of the new radio access technology of the fifth generation mobile communication system(5G),called new radio(NR).Many approaches have been investigated trying to improve the EE of wireless networks from different points of view:(i)network plan,(ii)renewable energy supply,(iii)system design.First,heterogeneous cellular networks(HetNets)in which macro cell served by macrocell base station(MBS)with higher power consumption and larger coverage is overlaid with small cells served by smallcell base stations(SBSs)with lower transmit power and smaller coverage,are promising network deployment for improving network EE.Second,energy harvesting wireless networks with power transfer is the other approach to improving EE by reducing the amount of traditional energy consumption.Third,radio resource allocation in terms of spectrum assignment and power allocation can be designed with the goal of maximizing the network EE.Building on the above observation and combining these three aspects,this project focuses on radio resource management and optimization for renewable energy powered HetNet aiming at improving the utilization ratio of the renewable energy and EE of the network.The research work and main contributions of this project are provided in the following three aspects.1.Cooperative Renewable Energy Management in Green HetNet.Powering the BSs in cellular networks with renewable energy is a promising approach to reducing the consumption of traditional on-grid brown energy and the carbon dioxide emissions and economic cost coming with it.However,the generation rate of renewable energy,e.g.,solar or wind,is changing in both time and spatial domains and the price of the battery used to store the extra renewable energy is relatively high compared with using traditional energy sources.Furthermore,the traffic loads at individual BSs also exhibit temporal and spatial diversities.Hence,appropriate renewable energy management strategies are important for the renewable energy powered wireless networks of providing stable energy supply and satisfying the demand of the traffic load.The author comprehensively considers the traffic loads and the cost of storing renewable energy and proposes the cooperative renewable energy management strategy for renewable energy powered HetNet.First,the optimal renewable energy storage strategy is proposed based on inventory theory to determine the optimal storage amount of renewable energy that can satisfy the traffic demand at the lowest inventory cost.In each time interval,BSs in the HetNet compute their optimal storage amount of renewable energy based on the proposed optimal renewable energy storage strategy and are divided into energy suppliers and demanders based on the difference of their actual amount of generated renewable energy and the optimal storage amount of renewable energy.Second,the renewable energy trading policy is proposed based on game theory to innovate the energy cooperation between energy suppliers and demanders.In the numerical simulations,two cases that respectively have high and low storage cost for renewable energy are studied and different optimal storage amounts of renewable energy are achieved by the proposed optimal renewable energy storage strategy.The numerical results show that the proposed renewable energy trading policy can achieve higher amount of trading energy with a lower price compared with the reference static energy cooperation policy.2.Energy and Spectrum Efficient Resource Allocation in Resource Cooperation Enabled Green HetNet.EE and spectral efficiency(SE)are two key pillars of 5G NR.Energy harvesting and transfer as a promising way to improving EE introduces new challenges to the guarantee of both EE and SE due to its natural causality.The author considers the green HetNet in which the MBSs and SBSs are powered by renewable energy and proposes resource sharing and allocation policies to simultaneously guarantee the EE and SE of the network.By introducing the concept of resource efficiency(RE)the EE and SE are represented as the unified metric RE and the problem of resource allocation with the goal of jointly optimizing EE and SE is formulated as a non-convex fractional programming problem.First,by using the Dinkelbach method and Lagrange dual decomposition,the author proposes a centralized resource allocation algorithm to solve the non-convex fractional programming problem.Considering the computational complexity and the signaling consumption during executing the centralized resource allocation algorithm,the author proposes a distributed resource allocation algorithm based on multi-agent reinforcement learning(MARL).Each BS interacts with other BSs in the HetNet,obtains the information of their renewable energy budget and available spectrum bandwidth,and makes subcarrier assignment and power allocation decisions to maximize RE.Simulation results demonstrate the convergence of the proposed centralized and distributed resource allocation algorisms.In comparison with the reference resource allocation schemes the proposed resource allocation algorithms can significantly improve the RE of the network.Furthermore,the proposed centralized resource allocation algorithm outperforms the proposed distributed algorithm in terms of network RE at the price of adding a control center and high complexity.3.Energy Efficient Resource Allocation in Energy Cooperation Enabled Green HetNet.Maximizing EE through radio resource allocation for renewable energy powered HetNet with energy sharing is investigated.The goal is to maximize the network EE,conquer the instability of renewable energy sources and guarantee the fairness of users during allocating resources.The objective function is defined as a sum weighted EE of all links in HetNet.The resource allocation problem in terms of subcarrier assignment,power allocation is formulated as a mixed combinational and non-convex optimization problem.The author proposes a centralized resource allocation algorithm by iteratively solving the subcarrier assignment and power allocation problems in which the power allocation problem is solved analytically solving the KKT conditions of the problem and a water-filling problem thereafter.Considering the complexity and the extra control center added for executing the centralized resource allocation algorithm,the author also proposes a distributed resource allocation algorithm with lower complexity based on MARL.The numerical results demonstrate the convergence of the proposed centralized and distributed resource allocation algorithms and show that both centralized and distributed algorithms outperform the existing reference algorithms in terms of the network EE.