Abstract: This paper presents an adaptive scheduling mode and algorithm based on tree grid. This mode can schedule fine-grained tasks and big jobs adaptively. Through the real-time detection of grid, an adaptive grid load-balancing algorithm(Optimization-based Priority-Completion Time Heuristic Algorithm for Task Allocation, OPCTHATA) for task allocation is proposed, which is based on load state, task execution time, task transmission time and task properties. The experiments show that the proposed algorithms for the scheduling problem have better performance than other algorithms on load-balancing and optimum makespan.

Key words: grid computing, task scheduling, task pool, heuristic algorithm

摘要： 提出一种基于树型计算网格的自适应调度算法，实现对小粒度独立任务和用户大作业的自适应最优调度。通过对网格环境的实时检测，给出了基于节点负载状况、节点任务执行时间、任务传输时间和任务特性的自适应调度算法，即基于最优任务分配方案的启发式任务调度算法。通过实验与其他调度算法的比较，证明了所提出的任务调度算法在负载平衡和最优跨度方面具有明显的优越性。

关键词: 网格计算, 任务调度, 任务池, 启发式算法

CLC Number: 

• TP393