基于均匀离散PSO 算法的多QoS 网格任务调度策略

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.

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

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.