基于平均时间的线程池尺寸自适应调整算法

摘要

为了实现自适应调整线程池尺寸，提高并发程序处理的运行速度、改善效率和降低系统的资源开销，提出了一种基于任务平均处理时间的线程池尺寸自适应调整算法。首先研究线程池的一些特征数据以及用户请求的任务时间周期，提出了任务平均处理时间的概念。然后研究了不同任务类型下，任务平均处理时间与线程池的尺寸的相关性，提出了一种自适应调整算法。实验结果表明，该算法能够自适应调整线程池尺寸到适当区域，有效地提高应用程序的整体性能。%To realize adaptive adjusting the thread pool size,improve the operation speed of concurrent processing program,efficiency and reduce the system resource spending,a thread pool size adaptive adjusting algorithm based on task average time is presented. Firstly,some characteristics of the thread pool and the task time period of users request are studied,then the concept of task average processing time is put forward. After that,the correlation between the task average processing time and thread pool size under the different types of tasks is analyzed. A kind of adaptive adjustment algorithm is put forward. The experimental results show that the algorithm can adaptively adjust thread pool size to the appropriate area,effectively improving the overall performance of application program.