云存储部署优化的进化算法设计

Design of evolutionary algorithm for the optimization of cloud storage deployment

针对云计算环境下的云存储部署规划优化问题,全文研究和提出了一种进化算法并进行了求解.首先基于对象存储方法,设计了三层架构云存储模型.在建立的模型中,将云存储部署抽象为多目标优化调度问题,针对现有粒子群优化算法或遗传算法在解决这一优化调度问题过程中出现的收敛速度及调度效率等方面的不足,将2种算法进行有效融合,设计了混合进化算法来进行求解,讨论了进化算法的优化设计过程.同时,在云计算仿真平台CloudSim上,对所提出的优化算法进行仿真实验,并对云存储中负载均衡等性能指标的改善程度进行了检测.结果表明,所提出的混合进化算法有效实现了云存储的优化部署任务,与其他进化算法相比,能对性能指标实现更优配置.

Aiming at the optimization of cloud storage deployment in cloud computing environment,an evolutionary algorithm is proposed.Firstly,a cloud storage model w ith three-tier structure is designed by using the object storage method.In this model,the cloud storage deployment is defined as a multi-objective optimization problem.This problem can be solved by using the traditional particle sw arm optimization（PSO） algorithm or the genetic algorithm（GA）.How ever,the convergence speed and the scheduling efficiency are not satisfactory.To overcome the above limitation,a hybrid evolutionary algorithm is presented by combining PSO and GA.The optimization design process of this evolutionary algorithm is analyzed.Moreover,the proposed algorithm is tested on the simulation platform CloudSim.Meanw hile,the improvement of main performance index such as load balance is evaluated.The experimental results show that the proposed algorithm efficiently improves the performance of cloud storage deployment.Compared w ith other approaches,this evolutionary algorithm can achieve better configuration of performance index.