云环境中面向服务软件的演化部署优化方法

发布时间：2018-07-05 11:09

本文选题：云计算 + 面向服务的软件　；参考：《中国科学:信息科学》2017年06期

【摘要】：针对现有的部署优化方法在求解云环境中面向服务软件的部署优化问题时,无法处理服务实例和虚拟机实例的伸缩以及无法保障求解质量等问题,本文提出了一种新的部署优化方法.该方法以提高面向服务软件的运行性能和降低运行成本为目标构建问题模型,并设计了一种基于遗传算法的MGA-DO算法对其进行求解.MGA-DO算法采用基于组的编码方式对软件的部署方案进行编码,然后结合基于组的单点交叉操作,实现了在优化过程中对服务实例和虚拟机实例的伸缩.此外,该算法引入现有的部署优化经验,设计了多种局部搜索规则,以进一步提高算法的求解性能.最后,一系列模拟实验表明,相比现有的算法,MGA-DO算法在求解所研究的问题时表现出了更好的性能.
[Abstract]:In order to solve the problem of service-oriented software deployment optimization in cloud environment, the existing deployment optimization methods can not deal with the scalability of service instances and virtual machine instances, and can not guarantee the quality of solution. In this paper, a new deployment optimization method is proposed. The purpose of this method is to improve the performance and reduce the running cost of the service-oriented software. A genetic algorithm-based MGA-DO algorithm is designed to solve it. MGA-DO algorithm uses group-based coding to encode the deployment scheme of the software, and then combines the single-point crossover operation based on group. In the process of optimization, the service instance and virtual machine instance are scalable. In addition, the algorithm introduces the existing experience of deployment optimization and designs a variety of local search rules to further improve the performance of the algorithm. Finally, a series of simulation experiments show that the MGA-DO algorithm has better performance in solving the studied problems than the existing algorithms.
【作者单位】：武汉大学软件工程国家重点实验室;武汉大学计算机学院;
【基金】：国家高技术研究发展计划(863)(批准号:2012AA011204) 国家自然科学基金(批准号:61373038、61672392)资助项目
【分类号】：TP18;TP302

【相似文献】