基于均匀离散PSO算法的多QoS网格任务调度策略  

Jobs scheduling policy for gird with multi-QoS constraints using uniform-design discrete particle swarm optimization

作　　者：蒲汛[1,2] 彭喜化[2] 于显平[2] 卢显良[1] 

机构地区：[1]电子科技大学计算机学院,成都410073 [2]西南大学计算机与信息科学学院,重庆400716

出　　处：《控制与决策》2013年第6期808-814,共7页Control and Decision

基　　金：国家重大专项项目(2011ZX03002-003-02);中央高校基本科研业务费专项项目(XDJK2012C020)

摘　　要：针对网格环境中多服务质量(QoS)约束条件下独立任务调度问题,提出一种融合配方均匀设计与离散粒子群优化算法(UDPSO)的任务调度策略,以实现对独立任务优化调度的快速生成.该算法采用类似DPSO算法的速度和位置更新方法,结合配方均匀设计,快速衡量各QoS约束条件的适应度,以产生分布均匀且较优的Pareto解集,最终为系统提供一组较优的任务调度方案.仿真实验表明,该算法更符合网格调度的复杂环境,能够得到较短的任务执行时间和较均衡的QoS保障.@@@@One of the key technologies to improve the efficiency of grid computing is to solve the independent job scheduling problem under the QoS constraints. Therefore, this paper designes an algorithm named uniform-design discrete particle swarm optimization(UDPSO) to find a sufficient number of uniformly distributed and representative Pareto optimal solution for the problem, which can make the performance of gird system to be more efficient. To solve the problem, the velocity and position of particles are refined, and a new method is used to optimize those two parameters. Then, the uniform design method is employed to get distributed Pareto front in the objective space efficiently. Finally, the global convergence of the algorithm is proved. Simulation results show that this algorithm is more effective and practical.

关 键 词：离散粒子群优化算法 PARETO最优 均匀设计 服务质量约束 任务分配 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]