Profit Maximization Method Based On Hybrid Power Supply For Cloud Data Centers

With the development of science and technology,smart devices are more and more popular,resulting in a large amount of data,the era of big data and 5G has also come.The huge amount of data generated every day will be processed by the cloud data centers,which brings about a problem of concern to cloud data centers,namely the energy consumption cost and revenue of cloud data center.Traditional cloud data centers purchase thermal power from the grid,which is not only expensive,but also not economical or environmental-friendly due to the large amount of pollution generated by thermal power,causing serious pollution to the environment.For these problems,this paper considers the green energy of wind and solar energy,and assumes that the green energy only has the one-time cost input of equipment.Therefore,this paper proposes a research scheme for profit maximization of cloud data centers based on hybrid power supply,giving priority to green energy provided by cloud data centers as far as possible,and using thermal power purchased from the grid when green energy supply is insufficient.In order to achieve cloud data centers as far as possible environmental-friendly.The main work and contributions of this paper are as follows:Firstly,this paper selects the task of three types,and for each type task design the corresponding service level agreement(SLA),each service level agreement not only include each type application both maximum delay time of income but also more than the compensation of delay time,more in line with the actual situation.Secondly,this paper selects three cloud data centers and uses their real load,wind,solar and electricity price data at the same time,which makes the experiment more convincing.The first is the generation and processing of load sequence.In this paper,load set published by google cloud data centers and through public figures released by the wind,solar and electricity price data are taken as the research objects.After preprocessing of the original data set,it is aggregated into time series with specified time slot.Thirdly,the bi-objective optimization model is constructed.In this paper,considering cloud data center for different application server,the request arrival rate,the influence of factors such as the loss rate of tasks,thus proposes a way of based on the hybrid power supply profit maximization method for cloud data center,because this paper deal with the objective function is a constraint function,so,by means of a point function method with the constrained bi-objective function,a bi-objective optimization problem into unconstrained to be solved.In addition,this paper also comprehensively considers the influence of cloud data center on different servers,task arrival rate,task loss rate and other factors,so as to propose a research scheme of cloud data center profit maximization method based on hybrid power supply.And the bi-objective optimization method is adopted to solve the model.Finally,the improved multi-objective optimization algorithm is used to solve the optimization problem.Compared with NSGA-II,SPEA2 and MOEA/D algorithms,the improved multi-objective optimization algorithm proposed in this paper is tested on the classic ZDT series test functions,and three performance indexes are used to measure,and good performance evaluation results are obtained.Based on the above method to solve the optimization model objective,make it reach a relative trade-off,and evaluate from the optimization results,operation time and the expansibility of the method and other indicators,that is to say,the cost optimization to the minimum and the revenue to the maximum,so as to achieve better results in profit.