摘要： 根据云计算环境中用户任务调度的不同需求，提出一种基于改进离散粒子群优化算法的任务调度策略，可实现在短时间内对云计算任务的相对较优调度。将用户费用与该任务的截止时间相结合，构建相对合理的用户优先级，以引导算法的适应度函数的偏好。引入重优化判断准则，在保证算法有能力跳出局部最优的同时保持解的多样性，最终求出满足用户优先级偏好的任务调度映射。仿真实验结果表明，该策略更符合云计算调度的复杂环境，能得到全局较优的任务调度方案。

关键词: 云计算, 任务调度, QoS约束, Hadoop架构, 离散粒子群优化, 用户优先级

Abstract: This paper proposes an Improved Discrete Particle of Swarm Optimization(IDPSO) to optimize the task scheduling problem of cloud computing with user priority level preferences in a short time. It combines the user priority and the task deadline to establish an appropriate task priority to guide the algorithm fitness function, employs a re-optimization criterion to ensure that the algorithm has the ability to jump out of local optima, and ultimately obtains task scheduling mapping with user priority preference. Simulation experimental results show that this algorithm is fit for cloud computing environment, it can gain overall optimal task scheduling scheme.

Key words: cloud computing, task scheduling, QoS constrains, Hadoop architecture, Discrete Particle of Swarm Optimization(DPSO), user priority level

中图分类号: 

• TP311