Energy Storage Scheduling And Wireless Resource Allocation For 5G Base Stations

In recent years,5G networks have been widely used due to their advantages such as low latency and high speed.However,the high energy consumption problem caused by the dense deployment of5 G base stations has greatly increased the costs of operators.Although current energy-saving solutions can effectively save base station energy consumption and operating costs,with the increasing complexity of 5G network architecture,there are still some shortcomings in existing energy-saving solutions.For example,failure to fully utilize the schedulable potential of base station resources,and the risk of reducing communication quality while saving energy.At the same time,the demand for wireless resources in 5G networks is gradually increasing.Traditional resource allocation strategies are difficult to adapt to increasingly complex network environments.Unreasonable allocation strategies can make it difficult to fully utilize wireless resources,reduce user experience,and also exacerbate the energy consumption of base stations.Therefore,this article proposes a 5G base station resource management method based on deep reinforcement learning,which combines5 G base stations with optical storage microgrids to balance operational costs and communication quality.The main job responsibilities are as follows:In response to the problem of high operating costs in the integrated system of 5G base station optical storage microgrid,this thesis proposes an energy storage scheduling scheme based on DeepQ-Network.Firstly,establish a microgrid integrated system model and its energy storage and scheduling model with 5G base stations as the load;Secondly,the energy storage scheduling problem in the 5G base station optical storage microgrid integrated system is described as a Markov decision process,and the energy storage scheduling model is used as a state set,the energy storage charging and discharging behavior is used as an action set,and the daily operating cost is used as a reward function;Next,historical data such as photovoltaic power generation,electricity price,and load are input into the neural network for training to obtain the optimal energy storage scheduling strategy;Finally,in order to verify the effectiveness and applicability of the proposed algorithm,energy storage scheduling charts and daily cost comparison charts for three typical regions under time of use electricity prices were provided,as well as daily cost comparison charts for commercial areas under real-time electricity prices and time of use electricity prices.According to the simulation results,compared to the situation without any regulation,the proposed scheme can save about 46.5% of operating costs and improve the efficiency of energy resource utilization of 5G base stations.In response to the challenge of balancing operational costs and communication quality in the integrated 5G base station optical storage microgrid system,this thesis proposes a wireless resource allocation scheme based on Deep-Q-Network.It combines the energy storage scheduling model and communication model as the state set,the energy storage charging and discharging behavior,the selection of base stations,and the adjustment of transmission power as the action set,and the ratio of transmission rate to transmission power as the energy efficiency,which is jointly optimized with operating costs.The simulation results show that the network performance of this algorithm is superior to Q-learning and greedy algorithms,while reducing the energy consumption of the base station.