Design Of Energy Scheduling Schemes In Smart-Grids For Powering Mobile Communication Systems

The explosive increase of mobile data traffic and base stations have raised people's concern about the enhancement of energy efficiency in mobile communication systems.By integrating re-newable energy resources and employing the energy scheduling technologies,smart grids are able to provide electricity to mobile communication systems in a more flexible and economical way and,therefore,attracted more and more attention.Existing researches have combined the smart grids with various technologies that are aimed to reduce energy consumption in mobile commu-nication systems so as to minimize the total operational cost and maximize the social welfare at the meantime.This thesis investigates the energy scheduling problems within smart-grids-powered mobile communication systems.We proposed effective energy scheduling algorithms to achieve d-ifferent optimization objectives in order to improve the energy efficiency of mobile communication systems.The research background and developing status are introduced at the beginning of the thesis.Researches on downlink transmission model in mobile communication systems,zero-forcing(ZF)beamforming scheme,Coordinated Multipoint(CoMP)transmission and content-centric cloud-based Random Access Network(CRAN),convex optimization theory,Alternating Direction of Multipliers Method(ADMM)and Penalty Dual Decomposition(PDD)method have been carried out before the design.This thesis summarizes the applications of the optimization algorithms,ADMM and PDD method,in scheduling energy.Advices on further exploration are given after analysis on the existing researches.On this base,we study the energy scheduling problems in micro-grids-powered mobile cellular networks and content-centric CRAN.We propose specific energy scheduling schemes to maximize the cost efficiency in the first system and to promote the fair allocation of energy resources in the second system.In this thesis,a micro-grid-powered mobile cellular network is studied.We come up with the cost efficiency metric and jointly optimize the transmit power,committed outputs of conventional generators,charging and discharging amounts of energy storage devices as well as the transac-tion amount with the main grid so as to maximize the cost efficiency while satisfying a series of constraints,such as the supply-demand balance constraint and the quality of service(QoS)con-straint.We transform the formulated non-convex linear fractional problem into a convex problem and propose the dual decomposition based cost efficiency maximization algorithm(DDBCEMA)and the low complexity ADMM-based algorithm(LCABA),both of which can be implemented in a parallel and distributed fashion.We compare the proposed algorithms with the existing cost minimization algorithm.Simulation results show that the proposed DDBCEMA and LCABA are able to minimize the total cost while maximizing the users' utility such that each unit of the cost has been made full use of to acquire the maximal cost efficiency.The content-centric CRAN is also studied in this thesis.We proposed the metric of green degree.We jointly optimize the dynamic clustering matrix,the beamforming vector,the shared energy between base stations and the transaction amount with the main grid in order to maximize the minimum green degree among all base stations while meeting the supply-demand balance constraint,the QoS constraint and etc.The proposed algorithm is able to allocate more energy resources to base stations that have to serve more mobile users;therefore,the problem of non-uniform distribution of renewable energy resources can be effectively addressed and the utilization of renewable energy can be promoted greatly.We propose the PDD-based energy scheduling algorithm to address the max-min problem.The proposed algorithm decomposes the complicated original problem,which has a non-convex objective function and a set of non-linearly coupled variables,into several easily solvable subproblems.Simulation confirms the fast convergence and the effectiveness of the proposed algorithm in promoting the fair allocation of renewable energy among base stations.