网格任务的脉冲响应模型与预测控制调度策略

摘要

The feedback control policy is employed to solve the grid task scheduling problem. According to the parallelism of tasks, a task impulse response model is developed to represent the relation between the task acceptance speed and the grid resource usage. In order to reduce the adverse effect of dynamic uncertainties on tasks processing, a predictive control strategy is used to adjust the task acceptance speed online; and an artificial immune algorithm is used to maximize the throughput of the grid. Results of simulation study show the validity of the proposed model and the effectiveness of the algorithm.%采用反馈控制策略,处理网格环境中的任务调度问题.利用任务并行度等内部结构信息,在任务接纳速度与网格计算资源使用量之间,建立了任务脉冲响应模型.采用预测控制策略对任务接纳速度进行在线调节,以消除网格动态不确定性因素对于任务执行的影响.采用人工免疫算法进行优化求解,最大化网格吞吐能力.仿真结果验证了模型的正确性和本文算法的有效性.