Research On Base Station Sleep Technology In Heterogeneous Cellular Networks

The 5th Generation mobile communication technology(5G)is a new mobile communication technology developed after the previous four generations of the communication technology.The Dense Heterogeneous Cellular Network(DHCN)has became an important technology to solve the surge of mobile users and mobile terminals in the 5G network environment.The deployment of large-scale low-power nodes and the random distribution of users lead to the increase of energy consumption and the waste of resources in the mobile communication system.The base station sleep,has received many attention and research as a key technology in the green communication advocated today.In the related research on the sleep technology of the base stations in the dense heterogeneous cellular network,there are still some problems in analyzing the dynamic changes of users and establishing the user association mechanism.In order to effectively reduce the energy consumption and improve the energy efficiency of the system in the scenario of the double-layer heterogeneous cellular network,two base station sleep algorithms are proposed in this thesis.Firstly,in order to reduce the energy consumption of the system,this thesis adopts the Femto Base Stations(FBSs)dynamic sleep algorithm based on the Semi-Markov Decision Process(SMDP).The FBS sleep process is modeled as a SMDP model.The number of users served by the base stations and the event that a user arrives or leaves are represented as the state of the system.The control of the FBS's state is represented as the action of the system.Then,the transition probability of each state transiting to other states can be obtained.The product of the power consumption and dwell time of the system after each state takes action is expressed as the revenue function of the system.The long-term return value function of the SMDP is further derived in this thesis,and the optimal sleep decision is obtained by the value iteration algorithm.The simulation results show that the proposed SMDP-based base station sleep algorithm can effectively reduce the energy consumption of the system.Secondly,in order to improve the energy efficiency of the system,this thesis adopts the base station sleep algorithm based on Sarsa learning.The algorithm comprehensively considers the energy efficiency of the uplink channel and the energy efficiency of the downlink channel.The number of the users served by the base station is represented as the state of the system.The control of the FBS's switch is represented as the action of the system,and the energy efficiency of the system after each state takes action is represented as the revenue function of the system.Through the Sarsa learning algorithm,the users in the system and the base station continuously perform information interaction,and the cumulative revenue is maximized to make the optimal base station sleep decision.In addition,this thesis proposes a user re-association mechanism to ensure the service continuity of the associated users of the sleep base stations and reduce the outage probability of the system users.Simulation results show that the base station sleep algorithm based on sarsa learning can effectively improve the energy efficiency of the system.