多云环境的虚拟应用网络部署决策方法研究

发布时间：2018-04-12 18:36

本文选题：多云 + 虚拟应用网络　；参考：《计算机科学》2017年S2期

【摘要】：多云环境下,用户能够基于虚拟应用技术与基础设施虚拟技术自由组合基础设施资源以部署虚拟应用网络,并且能够快速构建具有一定业务功能的分布式应用系统。鉴于现有多云部署决策方法,以及在部署描述及处理用户多目标需求方面的不足,首先给出了基于开放虚拟格式文档的虚拟应用网络部署描述方法;其次研究定义了基础设施资源组合的常见服务质量指标及多目标优化模型,并使用第二代非支配排序遗传算法(NSGA-II)和多目标粒子群优化(MOPSO)算法求解模型;最后给出了一种基于模糊决策的最终满意解确定方法。多次实验统计表明,两种算法均能在合理时间内实现较好收敛,NSGA-II适合2~3个目标的场景,而MOPSO能够用于更多目标场景,且均有更好表现;模糊决策所得最终解能够最佳匹配用户的目标偏好。
[Abstract]:In a cloudy environment, users can freely combine infrastructure resources based on virtual application technology and infrastructure virtual technology to deploy virtual application networks, and can quickly build distributed application systems with certain business functions.In view of the existing cloud deployment decision methods and the shortcomings in deployment description and user multi-objective requirements, a virtual application network deployment description method based on open virtual format document is proposed.Secondly, the common quality of service index and multi-objective optimization model of infrastructure resource combination are defined, and the second generation non-dominated sorting genetic algorithm (NSGA-II) and the multi-objective particle swarm optimization (MOPSO) algorithm are used to solve the model.Finally, a method for determining the final satisfactory solution based on fuzzy decision is presented.Many experiments show that both algorithms can achieve better convergent NSGA-II scenarios suitable for 2 ~ 3 targets in a reasonable time, while MOPSO can be used in more target scenarios and both have better performance.The final solution of fuzzy decision can best match the user's goal preference.
【作者单位】：解放军理工大学指挥信息系统学院;
【基金】：江苏省自然科学基金项目(BK20131069)资助
【分类号】：TP393.09

【相似文献】