应用GPD模型评估实时通信系统WCET

摘要 随着软件规模和硬件复杂性的不断增长，由确定性分析方法评估的最坏情况执行时间（WCET）越来越悲观，大大降低了系统的资源利用率.为提高实时系统的运行效率，应用极值统计方法评估软件的WCET，建立了WCET的广义Pareto分布统计模型.通过实验验证了该模型，证明了这种方法的可行性与有效性.
In design of real-time communication systems, the Worst-Case Execution Time (WCET) of software supplies pledge for the scheduling of the task, it is a creditable base that keeping the system running safely. With continuous growth of software scale and complexity of the hardware, the values of WCET estimated by determinate methods become more and more pessimistic, which greatly reduce the utilizing rate of the system resources . In order to improve the running efficiency of real time system, the article introduced a method of extreme value statistic in the estimation of WCET, and established the GPD model of WCET. The model was verified by experiments, the feasibility and the validity of this method was proven.