一种经典遗传算法下的异构云环境能效优化策略  

A Strategy of Energy-Efficency in Heterogeneous Cloud Computing Based on Classic Genetic Algorithm

作　　者：周航[1] 朱海[1,2] 齐迎春[1] 

机构地区：[1]周口师范学院计算机科学与技术学院,河南周口466001 [2]西安交通大学计算机系,陕西西安710049

出　　处：《信阳师范学院学报（自然科学版）》2013年第4期620-624,共5页Journal of Xinyang Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(U1204618;61103143);河南省高校科技创新人才支持项目(2012HASTII032)

摘　　要：从高能效的角度分析目前一些通用能耗模型,提出了一种适用于云数据中心实际应用场景的能效模型.根据此能效模型数学公式的分析,提出了相应的能效优化策略,并通过CloudSim仿真平台进行验证.结果表明,相比于Hadoop等云平台下传统的任务均分策略,应用经典遗传算法的能效优化策略能明显提高服务器集群的整体能效值.通过修改实验的相关参数,进一步验证了该能效优化策略的健壮性.By analyzing some general energy consumption model from the Energy-Effieency（EE） perspective, an EE model was presented for applying to Cloud-Computing data centre＇ s actual scenario. By analyzing the mathematical formula of the EE model, a corresponding strategy of EE optimization was presented and verified on the CloudSim Plat- form. The experimental results showed that the strategy based on the classic genetic algorithm can significantly improve the integral EE value of server cluster. Furthermore, the robustness of our strategy was also validated by modifying the relevant parameters of the experiment.

关 键 词：能效优化 云数据中心 异构性 

分 类 号：TP368.6[自动化与计算机技术—计算机系统结构] TP399[自动化与计算机技术—计算机科学与技术]