文章摘要
杨学佳,孙 浩,钱 健,张铁峰.基于改进NSGA-II算法的数据中心能量调度优化研究[J].电力需求侧管理,2025,27(4):64-70
基于改进NSGA-II算法的数据中心能量调度优化研究
Research on energy scheduling optimization for data center based on improved NSGA-II algorithm
投稿时间：2025-01-13  修订日期：2025-03-01
DOI：10. 3969 / j. issn. 1009-1831. 2025. 04. 010
中文关键词: 数据中心  多目标遗传算法  能量调度  低碳
英文关键词: data center  NSGA-II  energy scheduling  low-carbon
基金项目:国家自然科学基金（62273146）
作者单位
杨学佳 华北电力大学 电气与电子工程学院,河北 保定 071003 
孙 浩 内蒙古龙源新能源发展有限公司,呼和浩特 010000 
钱 健 内蒙古龙源新能源发展有限公司,呼和浩特 010000 
张铁峰 华北电力大学 电气与电子工程学院,河北 保定 071003；河北省电力物联网技术重点实验室（华北电力大学）,河北 保定 071003 
摘要点击次数: 624
全文下载次数: 536
中文摘要:
      为了满足数据中心低碳和经济的需求,建立以碳排放最少和综合供能成本最小为目标的数据中心能量调度优化模型,并利用改进快速和精英机制的多目标遗传算法（non-dominated sorting genetic algorithm-II,NSGA-II）对模型进行求解。首先介绍了数据中心综合能源系统优化调度框架,考虑数据中心的负载响应特性和设备能耗模型,建立了以经济和低碳为目标的数据中心能量调度优化模型。其次针对传统NSGA-II存在的Pareto解集分布不均匀,多样性较差的问题,提出一种改进NSGA-II算法。采用动态距离比较和精英保留方式选择个体,在保证优秀解的同时提高解的多样性。最后,基于某数据中心的能量调度算例验证了所提模型和方法在降低数据中心碳排放水平与综合能源系统供能成本方面的有效性。
英文摘要:
      To meet the low-carbon and economic demands of data centers, an energy scheduling optimization model for data centers with the objectives of minimizing carbon emissions and comprehensive energy costs is established. An improved non-dominated sorting genetic algorithm-II（NSGA-II）with enhanced speed and elite mechanism is proposed to solve the model. Firstly, the optimization scheduling framework for data center comprehensive energy systems is introduced, considering the load response characteristics of data centers and equipment energy consumption models, and an energy scheduling optimization model for data centers with economic and low-carbon objectives is established. Secondly, addressing the issue of uneven distribution of Pareto solution sets and poor diversity in traditional NSGA-II,an enhanced NSGA-II algorithm is proposed. It adopts dynamic distance comparison and elite retention selection of individuals to ensure both excellent solutions and improved diversity. Finally, through a case study of energy scheduling in a particular data center, the effectiveness of the model and method in reducing data center carbon emissions and comprehensive energy system costs is verified.
查看全文   查看/发表评论  下载PDF阅读器
关闭