Cloud Services Optimization Problem on Energy Utility Resource Allocation
作者单位:Computing CenterNortheastern University College of Information Science and EngineeringNortheastern University School of Economics and ManagementShenyang Aerospace University
会议名称:《第26届中国控制与决策会议》
会议日期:2014年
学科分类:08[工学] 080402[工学-测试计量技术及仪器] 0804[工学-仪器科学与技术]
基 金:supported by National Nature Science Foundation under Grant No.71171039 Fundamental Research Funds for the Central Universities under Grant No.N110204005,No.N120816001
关 键 词:Cloud Services Resource Allocation Energy Utility Particle Swarm Optimization
摘 要:In this paper, the cloud services optimization problem considering energy consumption cost is discussed. The queue model is presented for customer request service on data center. Since the server energy utility is based on the CPU core frequency, the novel trade-off optimization model between services revenue and energy loss cost is proposed in the paper, including allocating the dynamic CPU frequency of server. The HPSO algorithm is proposed for solving the mixed integer programming model. Then novel particle encoding substituting discrete variables and continuous variables efficiently improves solution quality. Finally, the dynamic penalty function method is discussed to convert the constraint optimization problem to non-constraint optimization problem.