摘要
云计算资源调度一直以来都是研究的热点,本文在云计算中引入粒子群算法,针对该算法局部收敛速度快,容易陷入局部最优值的缺点.本文提出了两个改进:一个是在粒子群种群寻找最优解中引入差分遗传算法,既可以发挥粒子群全局搜索快的优点,又可以发挥差分遗传算法局部搜索效率高的优点,将两种算法优点进行结合弥补粒子群算法不足;另一个是引入惩罚函数避免了粒子向无效的空间移动,节约了移动的成本.Cloudsim平台说明本文算法能够有效满足云计算资源分配,同时在任务完成时间,成本消耗方面都有了很大的提高,为云计算的资源分配提供了一种参考.
Resource scheduling in cloud computing has long been the focus of research. In this paper, particle swarm algorithm is introduced into the cloud computing and aiming at the shortcomings of this algorithm like fast local convergence and being easy to fall into local optimal value. Two improvements are made in this paper： one is to introduce differential genetic algorithm into particle swarm algorithm while finding the optimal solution, which can not only give play to particle swarm＇s advantage of quick global searching speed, but also give play to differential genetic algorithm＇s advantage of efficient local researching While combining the advantages of these two algorithms and making up for the deficiency of particle swarm algorithm. Another is to introduce penalty function so as to avoid particles moving towards void space and save costs of moving. Cloudsim platform shows that algorithm in this paper can effectively meet resource scheduling in cloud computing while having great improvement in reducing time of completing the task as well as consumption of costs so as to provide a reference for resource scheduling in cloud computing.
出处
《计算机系统应用》
2015年第10期136-141,共6页
Computer Systems & Applications
关键词
云计算
粒子群算法
差分遗传算法
惩罚函数
cloud computing
particle swarm algorithm
differential genetic algorithm
penalty function
