文献类型：期刊文献

中文题名：基于进化博弈模型的多层次云应用部署

英文题名：MULTI-LEVEL CLOUD APPLICATION DEPLOYMENT BASED ON EVOLUTIONARY GAME MODEL

作者：王亚利[1];文欣秀[2]

机构：[1]济源职业技术学院成教中心,河南济源459000;[2]华东理工大学信息科学与工程学院,上海200237

年份：2022

卷号：39

期号：9

起止页码：298

中文期刊名：计算机应用与软件

外文期刊名：Computer Applications and Software

收录：CSTPCD;;北大核心:【北大核心2020】；

基金：济源市市级重点实验室研究项目(2016-3)。

语种：中文

中文关键词：多层次云应用;进化博弈;进化稳定策略;应用部署;响应时间;Pareto支配

外文关键词：Multi-level cloud application;Evolutionary game;Evolutionary stable strategy;Application deployment;Response time;Poreto dominant

摘要：为了满足云应用部署的适应性和可扩展性,提出一种基于进化博弈模型的多层次云应用部署算法EGMAP。该算法可以确保每个云应用种群找到一种进化稳定部署策略。通过该博弈策略,在给定的系统负载和云资源可用性条件下,云应用可确定其最优的部署位置和相应资源分配。对算法稳定性进行了理论分析,证明应用种群状态可以收敛在部署策略的进化稳定策略点上,并证明了均衡解具有渐近稳定性。仿真实验结果表明,与基准算法相比,该算法在应用的响应时间、资源利用率、能耗指标上均表现出更好的性能。
In order to meet the adaptability and scalability of the cloud application deployment,this paper proposes a multi-level cloud application deployment algorithm based on evolutionary game model EGMAP.This algorithm could ensure to find an evolutionary stable deployment strategy for each cloud application population.Through this game strategy,for a given system workload and cloud resource availability,cloud applications can determine its optimal deployment location and the corresponding resource allocation.The stability of this algorithm were analyzed theoretically,and we proved that using the state of the population can converge to the evolutionary stable strategy point ESS on deployment strategy,and the equilibrium solution obtained was asymptotically stable.Simulation experimental results show that,compared with the baseline algorithms,this algorithm performs better on the response time of cloud applications,the resource utilization and the energy consumption of physical servers.