基于QPSO-SFLA改进算法的云环境资源调度研究

发布时间：2018-03-26 18:14

本文选题：云计算　切入点：资源调度　出处：《江西理工大学》2014年硕士论文

【摘要】：云计算是当前信息技术领域的热点话题之一,受到了工业界、学术界、政府和社会各界广泛关注。经过学术界与产业界的不断努力,云计算技术正逐渐从理论走向实践,各IT巨头纷纷加入到云计算的应用中来。云环境中的资源调度策略已成为云计算领域的一个重要研究方向。云计算调度策略的核心技术是将网络上众多的软硬件设备和资源,通过虚拟化技术整合成一个灵活的云处理系统,实现有效的监管。因此,对资源调度策略的研究一直是云计算领域的一个研究热点。本文在系统分析云计算的各种关键技术基础上,针对云环境下资源调度算法开展了重点研究,具体工作如下:(1)本文从资源管控策略、调配路径及调配算法等方面对云计算供应策略、云计算资源调度策略以及资源调度的性能指标、云计算的负载均衡技术等进行了全面分析研究。(2)本文探索了混洗蛙跳算法(SFLA)在云环境资源调度中的运用,在分析混洗蛙跳算法不足的基础上对种群选择策略进行了改进。改进后的混洗蛙跳算法(ISFLA)在种群初始化时引入轮盘赌随机选取策略,通过加强位置较优个体的适应度值提高了算法的收敛速度;在子种群重新混洗时引入元胞自动机策略,有效避免算法陷入局部最优。(3)将上述改进的SFLA算法与量子粒子群搜索策略结合,提出了基于量子粒子群局部搜索的新混洗蛙跳算法(QPSO-SFLA),从而提高了算法的局部搜索效率,加速算法的收敛。(4)在Cloud Sim平台上模拟云计算资源调度的过程及任务请求方式,对上述改进算法类进行了实验验证。将QPSO-SFLA算法、ISFLA算法及平台自带算法对比实验结果表明,QPSO-SFLA算法在云计算资源调度上具有明显的效率和成本优势。
[Abstract]:Cloud computing is one of the current hot topics in the field of information technology, which has received extensive attention from industry, academia, government and all walks of life. Through the continuous efforts of academia and industry, cloud computing technology is gradually moving from theory to practice. Various IT giants have joined the application of cloud computing. Resource scheduling strategy in cloud environment has become an important research direction in cloud computing field. The core technology of cloud computing scheduling strategy is to put many hardware and software devices and resources on the network. Through virtualization technology to integrate into a flexible cloud processing system to achieve effective regulation. Therefore, The research of resource scheduling strategy has been a hot topic in cloud computing field. Based on the systematic analysis of various key technologies of cloud computing, this paper focuses on resource scheduling algorithm in cloud environment. The specific work is as follows: (1) this paper analyzes the cloud computing supply strategy, cloud computing resource scheduling strategy and resource scheduling performance index from the aspects of resource control strategy, deployment path and deployment algorithm, etc. The load balancing technology of cloud computing is analyzed and studied comprehensively. 2) this paper explores the application of shuffled leapfrog algorithm (SFLAs) in cloud environment resource scheduling. Based on the analysis of the deficiency of the shuffled breaststroke algorithm, the population selection strategy is improved. The improved shuffling breaststroke algorithm (ISFLA) introduces the roulette random selection strategy when the population is initialized. The convergence rate of the algorithm is improved by strengthening the fitness value of the individuals with better position, and the cellular automata strategy is introduced when the subpopulation is remixed. The improved SFLA algorithm is combined with quantum particle swarm optimization (QPSO) strategy, and a new hybrid leapfrog algorithm (QPSO-SFLAA) based on QPSO local search is proposed, which improves the local search efficiency of the algorithm. Accelerate the convergence of the algorithm on the Cloud Sim platform to simulate the process of cloud computing resource scheduling and task request mode, The experimental results show that QPSO-SFLA algorithm has obvious efficiency and cost advantage in cloud computing resource scheduling.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.07;TP18

【共引文献】