摘要
大数据需要大量的云资源进行数据处理和分析,运行起来需要消耗较多的能源,在处理大数据的云环境中,资源数量和任务数量呈指数级增长,导致云数据中心的功耗增加。基于此,提出了一种基于强化学习的云环境下大数据能效策略模型,该模型利用DPSO和DQN的集成来更好地估计和校正数据维数缺陷。将所提出的模型与传统的DQN和负载感知算法加以比较。结果表明,随着任务数量的增加,所提模型在大数据处理方面的性能优于传统DQN和负载感知算法,为绿色云环境下的资源配置提供了一种节能方案。
Big data requires a large amount of cloud resources for data processing and analysis,which consumes a lot of energy to run.In the cloud environment where big data is processed,the number of resources and tasks increases exponentially,leading to an increase of power consumption in cloud data centers.Based on this,a reinforcement learning based big data energy efficiency strategy model in cloud environment is proposed,in which the integration of DPSO and DQN is utilized to better estimate and correct data dimensionality defects.The proposed model was compared with traditional DQN and load sensing algorithms.The results indicate that as the number of tasks increases,the proposed model outperforms traditional DQN and load sensing algorithms in big data processing,providing an energy⁃saving schedule for resource allocation in green cloud environments.
作者
余少锋
廖崇阳
马一宁
游锦鹏
YU Shaofeng;LIAO Chongyang;MA Yining;YOU Jinpeng(Information and Communication Branch of China Southern Power Grid Energy Storage Co.,Ltd.,Guangzhou 510000,China;Energy Development Research Institute,CSG,Guangzhou 510000,China)
出处
《电子设计工程》
2025年第1期142-145,共4页
Electronic Design Engineering
关键词
强化学习
能效策略
大数据
云环境
reinforcement learning
energy efficiency strategies
big data
cloud environment
