摘要
云计算环境下传统独立任务调度算法容易导致较高资源能耗或较大任务时间跨度.针对该问题,文中提出了两种能量感知的任务调度算法,并利用遗传算法并行化搜索合理调度方案.两种算法在搜索过程中,分别通过能耗时间归一和能耗时间双适应度方法定义适应度函数并进行个体选择.仿真结果表明,与单独考虑时间或能耗相比,这两种算法能够更有效地缩短任务执行时间跨度,降低资源能耗.
Traditional independent task scheduling algorithms in cloud computing may results in higher energy consumption of resources or longer makespan of tasks. This paper proposes two energy-aware task scheduling algorithms of cloud computing to consider makespan and the total energy consumption- They adopt the genetic algorithm to parallel find the reasonable scheduling scheme. These two algorithms use the methods of unify and double fitness to define the fitness function and select individuals. Compared with the algorithm considering only about time or energy consumption, the simulation results demonstrate the two algorithms can complete the tasks and reduce the energy consumption efficiently.
出处
《微电子学与计算机》
CSCD
北大核心
2012年第5期188-192,共5页
Microelectronics & Computer
基金
国家自然科学基金项目(60863003
61063042)
新疆大学博士科研启动基金(BS090153)
关键词
云计算
任务调度
遗传算法
能量感知
时间跨度
cloud computing
task scheduling
genetic algorithm
energy-awareness
makespan
