网格环境下复杂过程系统优化计算服务及任务调度策略

针对目前优化计算服务缺乏对一般用户的优化计算支持和灵活的资源共享方式,提出利用网格环境、支持网络计算资源以并行的、协同的方式提供面向服务的优化功能的计算服务系统,并设计了对系统性能具有重要影响的动态任务调度算法,从而使系统能够适应复杂过程系统优化计算以及网格环境的典型特征.相应地,系统采用多任务队列的任务管理方式和计算节点组的资源管理方式,为该调度策略提供实现基础.测试结果体现了基于该调度算法的优化计算服务系统的良好性能,并为设计在该系统下高效运行的并行算法提供了参考依据.

Optimization Service for Complex Process System in Grid Environment and Its Task Scheduling Strategy

In contrast to the existing optimization services lacking support to general optimization requests of users and flexible approaches to resource sharing,a computing service system is proposed to take advantage of the grid environment and enable computational resources in the network to provide service oriented optimization in parallel and cooperative mode.Dynamic task scheduling algorithm is designed to adapt the system to typical characters of complex process system optimization and grid environment.Accordingly,the system manages its tasks and resources through multi-task-queues and groups of computing nodes,respectively,which is the base of the realization of its scheduling strategy.Testing results have demonstrated good performance of the optimization service system with such scheduling algorithm,and provided reference for designing efficient parallel algorithms running in the system.