一种基于进化博弈理论的虚拟机部署算法  

A VIRTUAL MACHINE DEPLOYMENT ALGORITHM BASED ON EVOLUTIONARY GAME THEORY

作　　者：乐艺[1] Le Yi(Nanjing City Vocational College,Nanjing 210002,Jiangsu,China)

机构地区：[1]南京城市职业学院,江苏南京210002

出　　处：《计算机应用与软件》2021年第7期106-112,共7页Computer Applications and Software

基　　金：江苏省教育厅2017年高校“青蓝工程”项目。

摘　　要：为了实现云环境中自适应性和稳定性的应用执行与部署,提出一种基于进化博弈理论的多目标虚拟机部署算法。该算法可以确保每个云应用找到一种进化稳定部署策略:对于给定的系统负载和资源可用性,应用可确定其部署位置和相应资源分配。对算法的稳定性进行分析,证明了种群状态可收敛于部署策略的进化稳定策略ESS上,且得到的均衡解是渐近稳定的。通过三层架构的Web应用的仿真实验验证算法性能,结果表明,该算法在响应时间、资源利用率和功耗等指标上表现较优。In order to realize the cloud environment adaptability and stability of the application execution and deployment,this paper proposes a multi-objective virtual machine deployment algorithm based on evolutionary game theory.It could ensure that each cloud application finds an evolutionary stable deployment strategy.Under this policy,for a given system load and resource availability,cloud applications could determine its deployment location and corresponding resource allocation.The stability of the algorithm were analyzed,and we proved that the state of population could converge to the evolutionary stable strategy ESS on deployment strategy,and the equilibrium solution was asymptotically stable.The simulation experiments of Web applications with three-tier architecture prove the algorithm’s performance.The results show that the new algorithm performs better in response time,resource utilization and power consumption.

关 键 词：云计算 虚拟机部署 进化博弈 进化稳定策略 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]