中文摘要：

随着科学应用逐渐趋于数据密集型计算，为并行与分布式系统寻求高效的任务调度策略成了研究的热点问题。已有的可分任务调度模型均假设所有处理机都能100％的完成子任务的计算，即处理机在完成任务计算之前一直保持在线状态。实际上，并行与分布式系统中不同处理机的在线时间可能不同。若忽略处理机的在线时间，为其分配的任务量过大，则任务的完成时间可能超出处理机的下线时间，从而造成任务的计算无法按时完成。因此，为处理机分配任务时应充分考虑处理机下线时间的限制。为解决上述问题，该文提出了一种新的考虑处理机下线时间的可分任务调度优化模型，并设计了全局优化遗传算法求解该模型。最后，通过仿真实验结果验证了模型和算法的有效性。

英文摘要：

As scientific applications become more data intensive, finding an efficient scheduling strategy for massive computing in parallel and distributed systems has drawn increasingly attention. Most existing scheduling models assume that all processors can 100% finish computing, that is, they keep online during the completion of assigned workload fractions. In fact, in the real parallel and distributed environments, different processors have different off-line time. Therefore, off-line time constraints of processors should be taken into account before distributing of the workload fractions; otherwise, some processors may not be able to fish computing their assignments. To solve the above issue, this paper proposes an off-line time aware divisible-load scheduling model and designs an effective global optimization genetic algorithm to solve it. Finally, experimental results illustrate the effectiveness of the proposed model and the efficiency of the proposed algorithm.