摘要
针对云环境下任务调度易出现多目标冲突的问题,提出一种改进的基于猫群的多目标优化算法。该算法模拟猫的行为模式,采用基于线性混合比率的猫行为选择方式来提高全局搜索和局部寻优能力;并在迭代过程中结合任务完成时间和任务费用支出,引入一个可调节的多目标集成效用函数,实现了资源与任务的智能调度。实验结果表明,所提算法不仅求解质量高,且在求解速度和调度消耗方面均优于多目标遗传算法和多目标粒子群算法。
Considering the this paper proposes an optimized multi-objective conflict problem for job scheduling in cloud computing environment, multi-objective cat swarm algorithm (MO-CSO) based on cat behavior using the lin- ear mixture ratio selection method to improve the global search and local optimization ability. The algorithm combines makespan and total cost to propose an adjustable muhi-objective integrated utility function, which implements the intelligent dispatching of resources and task. The experimental results show that the optimized MO-CSO not only has high quality, but also are superior in solving speed and scheduling costs to the multi-objective genetic algorithm (MO-GA) and the multi-objective particle swarm optimization (MO-PSO) algorithm.
出处
《电子科技》
2016年第2期4-7,11,共5页
Electronic Science and Technology
基金
上海市自然科学基金资助项目(14ZR1440100)
上海市教委科研创新重点基金资助项目(14ZS167)
关键词
多目标猫群算法
线性混合比率
多目标集成效用函数
智能调度
multi-objective cat swarm algorithm
linear mixing ratio
a multi-objective integrated utility function
intelligent scheduling